[{"data":1,"prerenderedAt":126},["ShallowReactive",2],{"blog-archive-2026-07":3},[4,25,39,52,64,74,84,93,103,115],{"id":5,"slug":6,"body":7,"html":8,"title":9,"description":10,"category":11,"tags":12,"author":18,"date":19,"year":20,"month":21,"quarter":22,"status":23,"featured":24},"2026\u002F07\u002Findustry-applications\u002Foperational-risk-on-live-data","operational-risk-on-live-data","\nOperational risk functions are often stuck in a cycle: collect risk and control self-assessments in spreadsheets, consolidate them, report quarterly, repeat. By the time a report reaches the risk committee, the data is weeks old and the links between incidents, risks and controls have been lost along the way.\n\n## The connected model\n\nThe **risk management** family in the Atlas connects the objects risk teams already work with:\n\n- **Risk register:** risks by process, product and entity, with inherent and residual ratings.\n- **Controls:** mapped to risks, with owners and testing results.\n- **Key risk indicators:** thresholds and trends fed from source systems, not typed in.\n- **Incidents and loss events:** captured, classified, investigated and linked to the risks they reveal.\n- **Issues and actions:** remediation with owners, dates and verification.\n- **Assessments:** risk and control self-assessments run as workflows rather than spreadsheets.\n\nWhen these live in one application, questions like “which controls failed before this incident?” or “which risks have deteriorating KRIs and overdue actions?” become queries instead of projects.\n\n## Where AI helps\n\n- **Incident classification:** suggest a taxonomy category, root cause and the linked risks from the incident narrative.\n- **Pattern detection:** surface clusters of similar incidents across business units.\n- **Anomaly detection on KRIs:** flag unusual movements before they breach thresholds.\n- **Summarization:** draft committee papers from the underlying records, clearly marked as drafts.\n- **Assessment support:** pre-fill self-assessment answers from last cycle's evidence for owners to confirm or correct.\n\nRatings and risk acceptance stay with people. The application records when AI suggestions were used and whether they were accepted.\n\n## Who uses it\n\nRisk officers and operational risk teams, business-line risk champions, control owners, internal audit and executive management.\n\n## Integrations\n\nSource systems for KRI data, incident intake from ITSM and security tools, HR for ownership, finance for loss data, and the identity provider for role-based access to sensitive incidents.\n\n## Controls designed in\n\n- Four-eyes review of risk ratings\n- Evidence required for closing actions\n- Restricted visibility for sensitive investigations\n- A complete audit trail of rating changes\n\n## Why now\n\nSupervisors increasingly expect operational resilience: important business services mapped, impact tolerances set and scenarios tested. That is hard to evidence from spreadsheets. A connected risk application makes the mapping explicit and keeps it current.\n\n## First scope\n\nStart with incidents and KRIs for one business line, since that's where live data changes the conversation fastest, then extend to assessments. We'd scope it in a [Solution Definition Sprint](\u002Fservices\u002Fsolution-definition-sprint).\n\nSee [financial services](\u002Findustries\u002Ffinancial-services), explore the [Atlas](\u002Fatlas), or [bring us your risk workflow](\u002Fcontact).\n","\u003Cp>Operational risk functions are often stuck in a cycle: collect risk and control self-assessments in spreadsheets, consolidate them, report quarterly, repeat. By the time a report reaches the risk committee, the data is weeks old and the links between incidents, risks and controls have been lost along the way.\u003C\u002Fp>\n\u003Ch2>The connected model\u003C\u002Fh2>\n\u003Cp>The \u003Cstrong>risk management\u003C\u002Fstrong> family in the Atlas connects the objects risk teams already work with:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>Risk register:\u003C\u002Fstrong> risks by process, product and entity, with inherent and residual ratings.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Controls:\u003C\u002Fstrong> mapped to risks, with owners and testing results.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Key risk indicators:\u003C\u002Fstrong> thresholds and trends fed from source systems, not typed in.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Incidents and loss events:\u003C\u002Fstrong> captured, classified, investigated and linked to the risks they reveal.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Issues and actions:\u003C\u002Fstrong> remediation with owners, dates and verification.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Assessments:\u003C\u002Fstrong> risk and control self-assessments run as workflows rather than spreadsheets.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>When these live in one application, questions like “which controls failed before this incident?” or “which risks have deteriorating KRIs and overdue actions?” become queries instead of projects.\u003C\u002Fp>\n\u003Ch2>Where AI helps\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Cstrong>Incident classification:\u003C\u002Fstrong> suggest a taxonomy category, root cause and the linked risks from the incident narrative.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Pattern detection:\u003C\u002Fstrong> surface clusters of similar incidents across business units.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Anomaly detection on KRIs:\u003C\u002Fstrong> flag unusual movements before they breach thresholds.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Summarization:\u003C\u002Fstrong> draft committee papers from the underlying records, clearly marked as drafts.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Assessment support:\u003C\u002Fstrong> pre-fill self-assessment answers from last cycle's evidence for owners to confirm or correct.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>Ratings and risk acceptance stay with people. The application records when AI suggestions were used and whether they were accepted.\u003C\u002Fp>\n\u003Ch2>Who uses it\u003C\u002Fh2>\n\u003Cp>Risk officers and operational risk teams, business-line risk champions, control owners, internal audit and executive management.\u003C\u002Fp>\n\u003Ch2>Integrations\u003C\u002Fh2>\n\u003Cp>Source systems for KRI data, incident intake from ITSM and security tools, HR for ownership, finance for loss data, and the identity provider for role-based access to sensitive incidents.\u003C\u002Fp>\n\u003Ch2>Controls designed in\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>Four-eyes review of risk ratings\u003C\u002Fli>\n\u003Cli>Evidence required for closing actions\u003C\u002Fli>\n\u003Cli>Restricted visibility for sensitive investigations\u003C\u002Fli>\n\u003Cli>A complete audit trail of rating changes\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>Why now\u003C\u002Fh2>\n\u003Cp>Supervisors increasingly expect operational resilience: important business services mapped, impact tolerances set and scenarios tested. That is hard to evidence from spreadsheets. A connected risk application makes the mapping explicit and keeps it current.\u003C\u002Fp>\n\u003Ch2>First scope\u003C\u002Fh2>\n\u003Cp>Start with incidents and KRIs for one business line, since that's where live data changes the conversation fastest, then extend to assessments. We'd scope it in a \u003Ca href=\"\u002Fservices\u002Fsolution-definition-sprint\">Solution Definition Sprint\u003C\u002Fa>.\u003C\u002Fp>\n\u003Cp>See \u003Ca href=\"\u002Findustries\u002Ffinancial-services\">financial services\u003C\u002Fa>, explore the \u003Ca href=\"\u002Fatlas\">Atlas\u003C\u002Fa>, or \u003Ca href=\"\u002Fcontact\">bring us your risk workflow\u003C\u002Fa>.\u003C\u002Fp>\n","Operational risk management that runs on live data, not quarterly spreadsheets","Risk registers, KRIs, incidents and control testing as one connected application, with AI that helps risk teams see patterns earlier.","industry-applications",[13,14,15,16,17],"risk","financial-services","enterprise-operations","governance","evidence","fazezero-editorial","2026-07-30T00:00:00.000Z",2026,7,3,"published",false,{"id":26,"slug":27,"body":28,"html":29,"title":30,"description":31,"category":32,"tags":33,"author":18,"date":38,"year":20,"month":21,"quarter":22,"status":23,"featured":24},"2026\u002F07\u002Ffactory-and-architecture\u002Fidentity-tenancy-and-authorization-from-the-first-commit","identity-tenancy-and-authorization-from-the-first-commit","\nAsk what usually delays an AI application's go-live and the answer is rarely the AI. It's the security review. Specifically, the discovery that identity was bolted on, authorization checks are scattered through the code, and tenant isolation depends on every developer remembering a `where` clause.\n\nIn our factory, these three concerns are part of the architecture from the first commit. They're generated into the core of every application foundation.\n\n## Identity\n\nEvery foundation authenticates against an **enterprise identity provider** from day one: OpenID Connect or SAML to Entra ID, Okta, Ping, government SSO or a national identity platform. Local development uses the same flows with a local provider. There's no “temporary” username and password table waiting to be removed.\n\nService-to-service calls use their own identities, and AI agents act either as a service identity or on behalf of the invoking user, never with ambient superuser rights.\n\n## Authorization\n\nAuthorization is **explicit and central**:\n\n- Roles map to permissions in one place.\n- Permissions are checked in the application services layer, not in route handlers or UI code. Every entry point gets the same checks: the API, background jobs, agents and imports.\n- Consequential actions can require a second approver (maker\u002Fchecker) as a declared policy, not ad-hoc code.\n- Principals are re-evaluated per request, so a disabled user or a changed role takes effect immediately.\n\nThis is what makes a security review short. Reviewers read one policy map instead of hunting through handlers.\n\n## Multi-tenancy\n\nMany applications serve several tenants: business units, subsidiaries, government entities, partners, or customers of an SI. The foundations treat the **tenant as a first-class part of the data model**:\n\n- Tenant-owned records are stored and queried within the tenant's partition.\n- The tenant is taken from the authenticated principal, never from request data. A forged tenant ID in a payload is ignored.\n- Cross-tenant requests return “not found” rather than “forbidden”, so they don't leak existence.\n- Cross-tenant views, where legitimately needed, are separate and permissioned paths.\n\nIsolation is tested automatically: cross-tenant reads and writes are part of every foundation's test suite.\n\n## Why generate it instead of documenting it\n\nA guideline that says “always check the tenant” is followed until the deadline. A generated core that makes the unsafe path hard is followed every time. Because the same patterns run across the whole inventory, a weakness found in one place gets fixed in the factory and flows into every foundation that follows.\n\n## What still happens per customer\n\nThe integration with the customer's identity provider and directory, their role model and approval policies, their tenancy boundaries, and their security review. The rails are fixed, and the configuration is the customer's.\n\nMore on the rails: [Architecture](\u002Ffactory\u002Farchitecture) · [AI accelerates, architecture governs](\u002Fblog\u002Fai-accelerates-architecture-governs)\n","\u003Cp>Ask what usually delays an AI application's go-live and the answer is rarely the AI. It's the security review. Specifically, the discovery that identity was bolted on, authorization checks are scattered through the code, and tenant isolation depends on every developer remembering a \u003Ccode>where\u003C\u002Fcode> clause.\u003C\u002Fp>\n\u003Cp>In our factory, these three concerns are part of the architecture from the first commit. They're generated into the core of every application foundation.\u003C\u002Fp>\n\u003Ch2>Identity\u003C\u002Fh2>\n\u003Cp>Every foundation authenticates against an \u003Cstrong>enterprise identity provider\u003C\u002Fstrong> from day one: OpenID Connect or SAML to Entra ID, Okta, Ping, government SSO or a national identity platform. Local development uses the same flows with a local provider. There's no “temporary” username and password table waiting to be removed.\u003C\u002Fp>\n\u003Cp>Service-to-service calls use their own identities, and AI agents act either as a service identity or on behalf of the invoking user, never with ambient superuser rights.\u003C\u002Fp>\n\u003Ch2>Authorization\u003C\u002Fh2>\n\u003Cp>Authorization is \u003Cstrong>explicit and central\u003C\u002Fstrong>:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>Roles map to permissions in one place.\u003C\u002Fli>\n\u003Cli>Permissions are checked in the application services layer, not in route handlers or UI code. Every entry point gets the same checks: the API, background jobs, agents and imports.\u003C\u002Fli>\n\u003Cli>Consequential actions can require a second approver (maker\u002Fchecker) as a declared policy, not ad-hoc code.\u003C\u002Fli>\n\u003Cli>Principals are re-evaluated per request, so a disabled user or a changed role takes effect immediately.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>This is what makes a security review short. Reviewers read one policy map instead of hunting through handlers.\u003C\u002Fp>\n\u003Ch2>Multi-tenancy\u003C\u002Fh2>\n\u003Cp>Many applications serve several tenants: business units, subsidiaries, government entities, partners, or customers of an SI. The foundations treat the \u003Cstrong>tenant as a first-class part of the data model\u003C\u002Fstrong>:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>Tenant-owned records are stored and queried within the tenant's partition.\u003C\u002Fli>\n\u003Cli>The tenant is taken from the authenticated principal, never from request data. A forged tenant ID in a payload is ignored.\u003C\u002Fli>\n\u003Cli>Cross-tenant requests return “not found” rather than “forbidden”, so they don't leak existence.\u003C\u002Fli>\n\u003Cli>Cross-tenant views, where legitimately needed, are separate and permissioned paths.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>Isolation is tested automatically: cross-tenant reads and writes are part of every foundation's test suite.\u003C\u002Fp>\n\u003Ch2>Why generate it instead of documenting it\u003C\u002Fh2>\n\u003Cp>A guideline that says “always check the tenant” is followed until the deadline. A generated core that makes the unsafe path hard is followed every time. Because the same patterns run across the whole inventory, a weakness found in one place gets fixed in the factory and flows into every foundation that follows.\u003C\u002Fp>\n\u003Ch2>What still happens per customer\u003C\u002Fh2>\n\u003Cp>The integration with the customer's identity provider and directory, their role model and approval policies, their tenancy boundaries, and their security review. The rails are fixed, and the configuration is the customer's.\u003C\u002Fp>\n\u003Cp>More on the rails: \u003Ca href=\"\u002Ffactory\u002Farchitecture\">Architecture\u003C\u002Fa> · \u003Ca href=\"\u002Fblog\u002Fai-accelerates-architecture-governs\">AI accelerates, architecture governs\u003C\u002Fa>\u003C\u002Fp>\n","Identity, tenancy and authorization from the first commit","Why identity, multi-tenancy and authorization are generated into the core of every fazeZERO application foundation instead of added before go-live.","factory-and-architecture",[34,35,36,37],"identity","architecture","application-factory","api-first","2026-07-29T00:00:00.000Z",{"id":40,"slug":41,"body":42,"html":43,"title":44,"description":45,"category":11,"tags":46,"author":18,"date":51,"year":20,"month":21,"quarter":22,"status":23,"featured":24},"2026\u002F07\u002Findustry-applications\u002Fai-in-the-soc-triage-and-investigation","ai-in-the-soc-triage-and-investigation","\nSecurity operations centres don't lack alerts. They lack analyst time. Every tool in the stack produces detections, and many are duplicates, benign or low value. Real incidents compete for attention with noise, and analysts spend a large share of their day gathering context rather than making judgements.\n\n## What the application does\n\nThe **security operations** family in the Atlas focuses on the workflow between detection and response:\n\n1. **Ingest:** alerts from SIEM, EDR, email security, identity and cloud security tools, normalized into one model.\n2. **Enrich:** asset ownership, user context, threat intelligence and related alerts attached automatically.\n3. **Correlate:** group related alerts into a single investigation.\n4. **Triage:** prioritize by severity, asset criticality and confidence.\n5. **Investigate:** a case with a timeline, evidence, notes and tasks.\n6. **Respond:** response actions through the organization's tools, with approvals for high-impact steps.\n7. **Close and learn:** a disposition, lessons learned and tuning feedback to the detection owners.\n8. **Report:** metrics for SOC leadership and control evidence for audit.\n\n## Where AI helps\n\n- **Summarization:** a plain-language summary of what happened, affected assets and the evidence so far.\n- **Triage support:** a suggested priority and likely disposition, with the reasoning shown.\n- **Investigation assistance:** suggested next queries and pivots, and drafted incident timelines.\n- **Agentic enrichment:** bounded, read-only lookups across tools to assemble context before an analyst opens the case.\n- **Reporting:** draft incident reports and management summaries.\n\n## Guardrails that matter here\n\nSecurity is where uncontrolled automation does the most damage. The application enforces:\n\n- **Read-only by default.** Enrichment agents can look, not act.\n- **Human approval for containment.** Isolating hosts, disabling accounts and blocking traffic require an analyst, and a second approver for high-impact actions.\n- **Prompt-injection awareness.** Alert content is treated as untrusted data, never as instructions.\n- **A full audit trail** of every AI suggestion, every action and who approved it.\n\nWe cover the general pattern in [agentic automation with human checkpoints](\u002Fblog\u002Fagentic-automation-with-human-checkpoints).\n\n## Who uses it\n\nSOC analysts (tier 1 to 3), incident responders, SOC managers, CISOs, and control owners who need evidence for audits.\n\n## Integrations\n\nSIEM and log platforms, EDR\u002FXDR, identity providers, email security, cloud security posture tools, ticketing and ITSM, threat intelligence feeds, and asset inventories or CMDBs.\n\n## Measuring it honestly\n\nTrack time to triage, time to contain, the share of alerts closed as benign and analyst hours per incident. Agree the baseline first. Improvements should show up in your own metrics, not in vendor claims.\n\n## Where it applies\n\nEnterprise SOCs, managed security providers, financial institutions with regulatory incident-reporting obligations, and government security operations.\n\nExplore the [Atlas](\u002Fatlas), or [bring us your triage queue](\u002Fcontact).\n","\u003Cp>Security operations centres don't lack alerts. They lack analyst time. Every tool in the stack produces detections, and many are duplicates, benign or low value. Real incidents compete for attention with noise, and analysts spend a large share of their day gathering context rather than making judgements.\u003C\u002Fp>\n\u003Ch2>What the application does\u003C\u002Fh2>\n\u003Cp>The \u003Cstrong>security operations\u003C\u002Fstrong> family in the Atlas focuses on the workflow between detection and response:\u003C\u002Fp>\n\u003Col>\n\u003Cli>\u003Cstrong>Ingest:\u003C\u002Fstrong> alerts from SIEM, EDR, email security, identity and cloud security tools, normalized into one model.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Enrich:\u003C\u002Fstrong> asset ownership, user context, threat intelligence and related alerts attached automatically.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Correlate:\u003C\u002Fstrong> group related alerts into a single investigation.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Triage:\u003C\u002Fstrong> prioritize by severity, asset criticality and confidence.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Investigate:\u003C\u002Fstrong> a case with a timeline, evidence, notes and tasks.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Respond:\u003C\u002Fstrong> response actions through the organization's tools, with approvals for high-impact steps.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Close and learn:\u003C\u002Fstrong> a disposition, lessons learned and tuning feedback to the detection owners.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Report:\u003C\u002Fstrong> metrics for SOC leadership and control evidence for audit.\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Ch2>Where AI helps\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Cstrong>Summarization:\u003C\u002Fstrong> a plain-language summary of what happened, affected assets and the evidence so far.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Triage support:\u003C\u002Fstrong> a suggested priority and likely disposition, with the reasoning shown.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Investigation assistance:\u003C\u002Fstrong> suggested next queries and pivots, and drafted incident timelines.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Agentic enrichment:\u003C\u002Fstrong> bounded, read-only lookups across tools to assemble context before an analyst opens the case.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Reporting:\u003C\u002Fstrong> draft incident reports and management summaries.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>Guardrails that matter here\u003C\u002Fh2>\n\u003Cp>Security is where uncontrolled automation does the most damage. The application enforces:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>Read-only by default.\u003C\u002Fstrong> Enrichment agents can look, not act.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Human approval for containment.\u003C\u002Fstrong> Isolating hosts, disabling accounts and blocking traffic require an analyst, and a second approver for high-impact actions.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Prompt-injection awareness.\u003C\u002Fstrong> Alert content is treated as untrusted data, never as instructions.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>A full audit trail\u003C\u002Fstrong> of every AI suggestion, every action and who approved it.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>We cover the general pattern in \u003Ca href=\"\u002Fblog\u002Fagentic-automation-with-human-checkpoints\">agentic automation with human checkpoints\u003C\u002Fa>.\u003C\u002Fp>\n\u003Ch2>Who uses it\u003C\u002Fh2>\n\u003Cp>SOC analysts (tier 1 to 3), incident responders, SOC managers, CISOs, and control owners who need evidence for audits.\u003C\u002Fp>\n\u003Ch2>Integrations\u003C\u002Fh2>\n\u003Cp>SIEM and log platforms, EDR\u002FXDR, identity providers, email security, cloud security posture tools, ticketing and ITSM, threat intelligence feeds, and asset inventories or CMDBs.\u003C\u002Fp>\n\u003Ch2>Measuring it honestly\u003C\u002Fh2>\n\u003Cp>Track time to triage, time to contain, the share of alerts closed as benign and analyst hours per incident. Agree the baseline first. Improvements should show up in your own metrics, not in vendor claims.\u003C\u002Fp>\n\u003Ch2>Where it applies\u003C\u002Fh2>\n\u003Cp>Enterprise SOCs, managed security providers, financial institutions with regulatory incident-reporting obligations, and government security operations.\u003C\u002Fp>\n\u003Cp>Explore the \u003Ca href=\"\u002Fatlas\">Atlas\u003C\u002Fa>, or \u003Ca href=\"\u002Fcontact\">bring us your triage queue\u003C\u002Fa>.\u003C\u002Fp>\n","AI in the SOC: alert triage and investigation with evidence","Security operations applications that use AI to enrich, summarize and prioritize alerts while analysts keep the decisions and the evidence trail.",[47,48,17,49,50],"cybersecurity","case-management","human-in-the-loop","agents","2026-07-28T00:00:00.000Z",{"id":53,"slug":54,"body":55,"html":56,"title":57,"description":58,"category":59,"tags":60,"author":18,"date":63,"year":20,"month":21,"quarter":22,"status":23,"featured":24},"2026\u002F07\u002Fai-in-production\u002Fdocument-intelligence-in-regulated-workflows","document-intelligence-in-regulated-workflows","\nRegulated workflows run on documents: identity documents, company registries, financial statements, invoices, contracts, permits, medical referrals, supplier certificates, audit reports. Extracting data from them is among the most valuable uses of AI, and among the easiest to get subtly wrong.\n\nA demo extracts ten fields from a clean PDF perfectly. Production brings scans, photos, handwriting, multiple languages, unusual layouts and documents that are simply the wrong document.\n\n## The production pattern\n\n**1. Classify first.** Before extracting anything, determine what the document is. A bank statement sent where a trade licence was expected should be caught at the door.\n\n**2. Extract to a schema.** Every document type has a defined schema of fields, types and formats. The model's output is validated against it, and anything that doesn't conform is rejected.\n\n**3. Validate against rules and sources.** Cross-check extracted values: totals that should add up, dates that should be in order, registration numbers that should exist in a registry, names that should match the application.\n\n**4. Carry confidence and provenance.** Every extracted field records where it came from on the page and how confident the extraction is. Reviewers see the source next to the value.\n\n**5. Route by confidence and risk.** High-confidence, low-risk fields flow straight through. Low-confidence or high-risk fields go to a human review queue. The thresholds are business decisions, not model defaults.\n\n**6. Learn from corrections.** Every human correction is recorded and becomes evaluation data for the next model or prompt change.\n\n## Where it appears across the Atlas\n\nDocument intelligence isn't a product on its own. It's a capability inside many application families:\n\n- **Onboarding and KYC\u002FKYB:** identity and company documents\n- **Case management:** evidence submitted by applicants ([AI-assisted case management](\u002Fblog\u002Fai-assisted-case-management))\n- **Referrals and pre-authorization** in healthcare ([care coordination](\u002Fblog\u002Freferrals-and-care-coordination))\n- **Permits** in the built environment ([permitting and inspections](\u002Fblog\u002Fpermitting-and-inspections-for-the-built-environment))\n- **Supplier assurance:** SOC reports and certificates ([third-party risk](\u002Fblog\u002Fthird-party-and-supplier-risk-reviews))\n- **Finance:** remittances and statements ([reconciliation](\u002Fblog\u002Freconciliation-and-exception-workbenches))\n\n## Controls designed in\n\n- Original documents retained, unaltered, with hashes\n- Extracted values linked to their source location\n- Every human override recorded, with the reviewer and reason\n- Access to sensitive documents restricted by role and logged\n- The model provider and hosting chosen to meet data-residency requirements\n\n## Measuring it honestly\n\nField-level accuracy on a held-out test set per document type, straight-through processing rate, review queue volume and correction rate. Agree the thresholds with the business and compliance owners before launch. See [evaluation and guardrails](\u002Fblog\u002Fevaluation-and-guardrails-before-production).\n\n## Arabic and bilingual documents\n\nIn the GCC, many documents are Arabic, English or both, and include stamps, signatures and handwriting. Test sets must reflect that mix from day one. Performance on English samples says little about performance on the documents you'll actually receive.\n\n[Bring us the document types](\u002Fcontact) that slow your workflow down.\n","\u003Cp>Regulated workflows run on documents: identity documents, company registries, financial statements, invoices, contracts, permits, medical referrals, supplier certificates, audit reports. Extracting data from them is among the most valuable uses of AI, and among the easiest to get subtly wrong.\u003C\u002Fp>\n\u003Cp>A demo extracts ten fields from a clean PDF perfectly. Production brings scans, photos, handwriting, multiple languages, unusual layouts and documents that are simply the wrong document.\u003C\u002Fp>\n\u003Ch2>The production pattern\u003C\u002Fh2>\n\u003Cp>\u003Cstrong>1. Classify first.\u003C\u002Fstrong> Before extracting anything, determine what the document is. A bank statement sent where a trade licence was expected should be caught at the door.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>2. Extract to a schema.\u003C\u002Fstrong> Every document type has a defined schema of fields, types and formats. The model's output is validated against it, and anything that doesn't conform is rejected.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>3. Validate against rules and sources.\u003C\u002Fstrong> Cross-check extracted values: totals that should add up, dates that should be in order, registration numbers that should exist in a registry, names that should match the application.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>4. Carry confidence and provenance.\u003C\u002Fstrong> Every extracted field records where it came from on the page and how confident the extraction is. Reviewers see the source next to the value.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>5. Route by confidence and risk.\u003C\u002Fstrong> High-confidence, low-risk fields flow straight through. Low-confidence or high-risk fields go to a human review queue. The thresholds are business decisions, not model defaults.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>6. Learn from corrections.\u003C\u002Fstrong> Every human correction is recorded and becomes evaluation data for the next model or prompt change.\u003C\u002Fp>\n\u003Ch2>Where it appears across the Atlas\u003C\u002Fh2>\n\u003Cp>Document intelligence isn't a product on its own. It's a capability inside many application families:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>Onboarding and KYC\u002FKYB:\u003C\u002Fstrong> identity and company documents\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Case management:\u003C\u002Fstrong> evidence submitted by applicants (\u003Ca href=\"\u002Fblog\u002Fai-assisted-case-management\">AI-assisted case management\u003C\u002Fa>)\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Referrals and pre-authorization\u003C\u002Fstrong> in healthcare (\u003Ca href=\"\u002Fblog\u002Freferrals-and-care-coordination\">care coordination\u003C\u002Fa>)\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Permits\u003C\u002Fstrong> in the built environment (\u003Ca href=\"\u002Fblog\u002Fpermitting-and-inspections-for-the-built-environment\">permitting and inspections\u003C\u002Fa>)\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Supplier assurance:\u003C\u002Fstrong> SOC reports and certificates (\u003Ca href=\"\u002Fblog\u002Fthird-party-and-supplier-risk-reviews\">third-party risk\u003C\u002Fa>)\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Finance:\u003C\u002Fstrong> remittances and statements (\u003Ca href=\"\u002Fblog\u002Freconciliation-and-exception-workbenches\">reconciliation\u003C\u002Fa>)\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>Controls designed in\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>Original documents retained, unaltered, with hashes\u003C\u002Fli>\n\u003Cli>Extracted values linked to their source location\u003C\u002Fli>\n\u003Cli>Every human override recorded, with the reviewer and reason\u003C\u002Fli>\n\u003Cli>Access to sensitive documents restricted by role and logged\u003C\u002Fli>\n\u003Cli>The model provider and hosting chosen to meet data-residency requirements\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>Measuring it honestly\u003C\u002Fh2>\n\u003Cp>Field-level accuracy on a held-out test set per document type, straight-through processing rate, review queue volume and correction rate. Agree the thresholds with the business and compliance owners before launch. See \u003Ca href=\"\u002Fblog\u002Fevaluation-and-guardrails-before-production\">evaluation and guardrails\u003C\u002Fa>.\u003C\u002Fp>\n\u003Ch2>Arabic and bilingual documents\u003C\u002Fh2>\n\u003Cp>In the GCC, many documents are Arabic, English or both, and include stamps, signatures and handwriting. Test sets must reflect that mix from day one. Performance on English samples says little about performance on the documents you'll actually receive.\u003C\u002Fp>\n\u003Cp>\u003Ca href=\"\u002Fcontact\">Bring us the document types\u003C\u002Fa> that slow your workflow down.\u003C\u002Fp>\n","Document intelligence in regulated workflows: extraction with verification","Extracting data from documents with AI is easy to demo and hard to trust. How to build extraction with validation, confidence and human review.","ai-in-production",[61,49,17,62],"document-intelligence","production","2026-07-23T00:00:00.000Z",{"id":65,"slug":66,"body":67,"html":68,"title":69,"description":70,"category":11,"tags":71,"author":18,"date":73,"year":20,"month":21,"quarter":22,"status":23,"featured":24},"2026\u002F07\u002Findustry-applications\u002Freconciliation-and-exception-workbenches","reconciliation-and-exception-workbenches","\nFew finance processes consume as much skilled time as reconciliation. Statements, ledgers, sub-ledgers, payment files and counterparty reports all have to agree, and when they don't, someone investigates. At month-end that “someone” is usually a team working in spreadsheets.\n\nIt is also one of the most practical places to apply AI, because the work is repetitive, the data is structured, the exceptions follow patterns and the outcome is verifiable.\n\n## What the application does\n\nThe **financial operations** family in the Atlas includes reconciliation workbench foundations built around five workflows:\n\n1. **Ingest.** Pull statements, ledger extracts and payment files through adapters, and normalize them into a common model.\n2. **Match.** Rule-based matching first (exact, tolerance, many-to-one), then suggested matches for what is left.\n3. **Investigate breaks.** Unmatched items become exceptions in a queue, with ageing, ownership and priority.\n4. **Resolve and approve.** Adjustments and write-offs go through maker\u002Fchecker approval, with the reason recorded.\n5. **Close and evidence.** Reconciliation sign-off with a full history, ready for audit.\n\n## Where AI helps, and where it doesn't\n\n**It helps with:**\n\n- suggesting matches for items that rules can't pair, with a confidence score and the reasoning shown\n- classifying breaks by likely cause (timing, fees, FX, duplicates, missing entries)\n- summarizing an exception's history for whoever picks it up\n- extracting data from unstructured remittance advice and statements\n- spotting anomalies such as unusual break volumes or recurring counterparty issues\n\n**It doesn't:**\n\n- post adjustments on its own\n- approve write-offs\n- change matching rules without review\n\nDeterministic rules stay in charge of the ledger. AI shortens the path to a human decision.\n\n## Who uses it\n\nFinance analysts and operations controllers do the daily work. Treasury managers need cash visibility. Controllers and CFO offices need the close. Internal audit needs the evidence.\n\n## Integrations\n\nERP general ledgers, banking APIs and statement formats (including ISO 20022 camt messages), payment hubs, card processors and, in digital-asset operations, custody and wallet balances. See [stablecoin settlement operations](\u002Fblog\u002Foperating-stablecoin-settlement).\n\n## Controls designed in\n\n- Segregation between preparer and approver\n- Thresholds that force a second approval on large adjustments\n- Immutable history of matches, unmatches and overrides\n- Ageing and escalation rules for unresolved breaks\n\n## Measuring success honestly\n\nThe metrics that matter are auto-match rate, exception ageing, time to close and the number of manual adjustments. We agree baselines during the [Solution Definition Sprint](\u002Fservices\u002Fsolution-definition-sprint), so success is measured against your numbers, not a vendor's brochure.\n\n## Where it applies\n\nBanks, payment companies, insurers, corporate treasury and shared-service centres, and digital-asset operators reconciling on-chain and off-chain records.\n\nSee [financial services](\u002Findustries\u002Ffinancial-services) or [bring us your reconciliation](\u002Fcontact).\n","\u003Cp>Few finance processes consume as much skilled time as reconciliation. Statements, ledgers, sub-ledgers, payment files and counterparty reports all have to agree, and when they don't, someone investigates. At month-end that “someone” is usually a team working in spreadsheets.\u003C\u002Fp>\n\u003Cp>It is also one of the most practical places to apply AI, because the work is repetitive, the data is structured, the exceptions follow patterns and the outcome is verifiable.\u003C\u002Fp>\n\u003Ch2>What the application does\u003C\u002Fh2>\n\u003Cp>The \u003Cstrong>financial operations\u003C\u002Fstrong> family in the Atlas includes reconciliation workbench foundations built around five workflows:\u003C\u002Fp>\n\u003Col>\n\u003Cli>\u003Cstrong>Ingest.\u003C\u002Fstrong> Pull statements, ledger extracts and payment files through adapters, and normalize them into a common model.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Match.\u003C\u002Fstrong> Rule-based matching first (exact, tolerance, many-to-one), then suggested matches for what is left.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Investigate breaks.\u003C\u002Fstrong> Unmatched items become exceptions in a queue, with ageing, ownership and priority.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Resolve and approve.\u003C\u002Fstrong> Adjustments and write-offs go through maker\u002Fchecker approval, with the reason recorded.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Close and evidence.\u003C\u002Fstrong> Reconciliation sign-off with a full history, ready for audit.\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Ch2>Where AI helps, and where it doesn't\u003C\u002Fh2>\n\u003Cp>\u003Cstrong>It helps with:\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cul>\n\u003Cli>suggesting matches for items that rules can't pair, with a confidence score and the reasoning shown\u003C\u002Fli>\n\u003Cli>classifying breaks by likely cause (timing, fees, FX, duplicates, missing entries)\u003C\u002Fli>\n\u003Cli>summarizing an exception's history for whoever picks it up\u003C\u002Fli>\n\u003Cli>extracting data from unstructured remittance advice and statements\u003C\u002Fli>\n\u003Cli>spotting anomalies such as unusual break volumes or recurring counterparty issues\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>\u003Cstrong>It doesn't:\u003C\u002Fstrong>\u003C\u002Fp>\n\u003Cul>\n\u003Cli>post adjustments on its own\u003C\u002Fli>\n\u003Cli>approve write-offs\u003C\u002Fli>\n\u003Cli>change matching rules without review\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>Deterministic rules stay in charge of the ledger. AI shortens the path to a human decision.\u003C\u002Fp>\n\u003Ch2>Who uses it\u003C\u002Fh2>\n\u003Cp>Finance analysts and operations controllers do the daily work. Treasury managers need cash visibility. Controllers and CFO offices need the close. Internal audit needs the evidence.\u003C\u002Fp>\n\u003Ch2>Integrations\u003C\u002Fh2>\n\u003Cp>ERP general ledgers, banking APIs and statement formats (including ISO 20022 camt messages), payment hubs, card processors and, in digital-asset operations, custody and wallet balances. See \u003Ca href=\"\u002Fblog\u002Foperating-stablecoin-settlement\">stablecoin settlement operations\u003C\u002Fa>.\u003C\u002Fp>\n\u003Ch2>Controls designed in\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>Segregation between preparer and approver\u003C\u002Fli>\n\u003Cli>Thresholds that force a second approval on large adjustments\u003C\u002Fli>\n\u003Cli>Immutable history of matches, unmatches and overrides\u003C\u002Fli>\n\u003Cli>Ageing and escalation rules for unresolved breaks\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>Measuring success honestly\u003C\u002Fh2>\n\u003Cp>The metrics that matter are auto-match rate, exception ageing, time to close and the number of manual adjustments. We agree baselines during the \u003Ca href=\"\u002Fservices\u002Fsolution-definition-sprint\">Solution Definition Sprint\u003C\u002Fa>, so success is measured against your numbers, not a vendor's brochure.\u003C\u002Fp>\n\u003Ch2>Where it applies\u003C\u002Fh2>\n\u003Cp>Banks, payment companies, insurers, corporate treasury and shared-service centres, and digital-asset operators reconciling on-chain and off-chain records.\u003C\u002Fp>\n\u003Cp>See \u003Ca href=\"\u002Findustries\u002Ffinancial-services\">financial services\u003C\u002Fa> or \u003Ca href=\"\u002Fcontact\">bring us your reconciliation\u003C\u002Fa>.\u003C\u002Fp>\n","Reconciliation and exception workbenches: where finance AI earns its keep","Why transaction and ledger reconciliation is one of the most practical AI applications in finance: matching, break investigation and evidence.",[72,14,15,61],"reconciliation","2026-07-21T00:00:00.000Z",{"id":75,"slug":76,"body":77,"html":78,"title":79,"description":80,"category":11,"tags":81,"author":18,"date":83,"year":20,"month":21,"quarter":22,"status":23,"featured":24},"2026\u002F07\u002Findustry-applications\u002Fai-assisted-case-management","ai-assisted-case-management","\nCase management is everywhere once you look for it: benefit applications, licensing requests, complaints, investigations, customer disputes, employee cases, service requests. The shape is the same each time. Something arrives, it's triaged, someone works it, a decision is made and it may be appealed. Backlogs grow when intake outpaces the people who decide.\n\nThat common shape is why case management is one of the most reusable application families in the Atlas, and one of the best places to apply AI safely.\n\n## The core workflow\n\n1. **Intake:** cases arrive through portals, email, APIs or other systems, with documents attached.\n2. **Triage:** each case is classified by type, urgency and complexity, and routed to the right queue.\n3. **Assignment:** workload-aware allocation to case workers, with skills and conflicts respected.\n4. **Work:** information requests, internal consultations, notes and deadlines.\n5. **Decision:** a structured decision with its rationale, approved where policy requires.\n6. **Communication:** notifications and letters to the applicant or customer.\n7. **Appeal or reopen:** a linked case with its full history.\n8. **Reporting:** backlog, ageing, service levels and outcomes.\n\n## Where AI helps\n\n- **Document intelligence:** extract fields from submitted documents and check completeness before a case reaches a person.\n- **Classification and routing:** suggest case type and priority, with the suggestion recorded.\n- **Case summaries:** a short, current summary at the top of every case, so a new case worker doesn't reread forty pages.\n- **Similar-case retrieval:** find precedents and relevant policy passages with citations.\n- **Drafting:** propose decision letters and information requests for the case worker to edit.\n\n## Where it must not\n\nAI never makes the decision in consequential cases. It doesn't deny, approve or close on its own. The workflow puts human checkpoints at every decision, records who decided, and keeps AI-generated text visibly marked until a person accepts it. In the public sector, this is about legitimacy as much as risk: citizens are entitled to an accountable decision-maker.\n\n## Controls designed in\n\n- Role-based access to sensitive case data\n- Conflict-of-interest checks on assignment\n- A complete audit history of every change, view and decision\n- Retention and disclosure rules configured per case type\n\n## Integrations\n\nCitizen or customer portals, national identity and SSO, document management, CRM or registry systems, payment systems for fees, and messaging services.\n\n## Where it applies\n\nGovernment and public services, financial services complaints and disputes, insurance claims triage, HR case management and enterprise service teams. The foundation is the same, and the domain vocabulary and policies are configured.\n\n## First scope\n\nOne case type with a real backlog. Measure time to first touch, time to decision and backlog ageing before and after. Scope it in a [Solution Definition Sprint](\u002Fservices\u002Fsolution-definition-sprint).\n\nSee [government and public sector](\u002Findustries\u002Fgovernment-public-sector), explore the [Atlas](\u002Fatlas), or [bring us your backlog](\u002Fcontact).\n","\u003Cp>Case management is everywhere once you look for it: benefit applications, licensing requests, complaints, investigations, customer disputes, employee cases, service requests. The shape is the same each time. Something arrives, it's triaged, someone works it, a decision is made and it may be appealed. Backlogs grow when intake outpaces the people who decide.\u003C\u002Fp>\n\u003Cp>That common shape is why case management is one of the most reusable application families in the Atlas, and one of the best places to apply AI safely.\u003C\u002Fp>\n\u003Ch2>The core workflow\u003C\u002Fh2>\n\u003Col>\n\u003Cli>\u003Cstrong>Intake:\u003C\u002Fstrong> cases arrive through portals, email, APIs or other systems, with documents attached.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Triage:\u003C\u002Fstrong> each case is classified by type, urgency and complexity, and routed to the right queue.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Assignment:\u003C\u002Fstrong> workload-aware allocation to case workers, with skills and conflicts respected.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Work:\u003C\u002Fstrong> information requests, internal consultations, notes and deadlines.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Decision:\u003C\u002Fstrong> a structured decision with its rationale, approved where policy requires.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Communication:\u003C\u002Fstrong> notifications and letters to the applicant or customer.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Appeal or reopen:\u003C\u002Fstrong> a linked case with its full history.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Reporting:\u003C\u002Fstrong> backlog, ageing, service levels and outcomes.\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Ch2>Where AI helps\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Cstrong>Document intelligence:\u003C\u002Fstrong> extract fields from submitted documents and check completeness before a case reaches a person.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Classification and routing:\u003C\u002Fstrong> suggest case type and priority, with the suggestion recorded.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Case summaries:\u003C\u002Fstrong> a short, current summary at the top of every case, so a new case worker doesn't reread forty pages.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Similar-case retrieval:\u003C\u002Fstrong> find precedents and relevant policy passages with citations.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Drafting:\u003C\u002Fstrong> propose decision letters and information requests for the case worker to edit.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>Where it must not\u003C\u002Fh2>\n\u003Cp>AI never makes the decision in consequential cases. It doesn't deny, approve or close on its own. The workflow puts human checkpoints at every decision, records who decided, and keeps AI-generated text visibly marked until a person accepts it. In the public sector, this is about legitimacy as much as risk: citizens are entitled to an accountable decision-maker.\u003C\u002Fp>\n\u003Ch2>Controls designed in\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>Role-based access to sensitive case data\u003C\u002Fli>\n\u003Cli>Conflict-of-interest checks on assignment\u003C\u002Fli>\n\u003Cli>A complete audit history of every change, view and decision\u003C\u002Fli>\n\u003Cli>Retention and disclosure rules configured per case type\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>Integrations\u003C\u002Fh2>\n\u003Cp>Citizen or customer portals, national identity and SSO, document management, CRM or registry systems, payment systems for fees, and messaging services.\u003C\u002Fp>\n\u003Ch2>Where it applies\u003C\u002Fh2>\n\u003Cp>Government and public services, financial services complaints and disputes, insurance claims triage, HR case management and enterprise service teams. The foundation is the same, and the domain vocabulary and policies are configured.\u003C\u002Fp>\n\u003Ch2>First scope\u003C\u002Fh2>\n\u003Cp>One case type with a real backlog. Measure time to first touch, time to decision and backlog ageing before and after. Scope it in a \u003Ca href=\"\u002Fservices\u002Fsolution-definition-sprint\">Solution Definition Sprint\u003C\u002Fa>.\u003C\u002Fp>\n\u003Cp>See \u003Ca href=\"\u002Findustries\u002Fgovernment-public-sector\">government and public sector\u003C\u002Fa>, explore the \u003Ca href=\"\u002Fatlas\">Atlas\u003C\u002Fa>, or \u003Ca href=\"\u002Fcontact\">bring us your backlog\u003C\u002Fa>.\u003C\u002Fp>\n","AI-assisted case management: summaries, triage and human decisions","Case management across government services and enterprise operations: intake, triage, assignment, decisions and appeals, with AI assisting.",[48,82,15,49,61],"government","2026-07-16T00:00:00.000Z",{"id":85,"slug":86,"body":87,"html":88,"title":89,"description":90,"category":59,"tags":91,"author":18,"date":92,"year":20,"month":21,"quarter":22,"status":23,"featured":24},"2026\u002F07\u002Fai-in-production\u002Fagentic-automation-with-human-checkpoints","agentic-automation-with-human-checkpoints","\nAgents, meaning AI systems that plan and take multi-step actions with tools, are the most exciting and the most dangerous AI capability in the enterprise. An agent that gathers context from five systems before an analyst opens a case saves real time. An agent that closes accounts, moves money or emails customers on its own is a governance incident waiting to happen.\n\nThe answer isn't to avoid agents. It's to put them inside a workflow with **checkpoints**.\n\n## Design principles\n\n**1. Bounded tools.** An agent can only call tools that the application explicitly exposes to it, each with a narrow purpose and validated inputs. No general shell, no arbitrary API access.\n\n**2. Read before write.** Most value comes from read-only work: gathering context, correlating records, drafting. Make read-only the default and treat every write as a separate, higher-risk capability.\n\n**3. Explicit approval for consequential actions.** Anything that changes a record of consequence, contacts a customer, moves value or changes access requires a person to approve. Some actions require two people.\n\n**4. Identity and least privilege.** The agent acts with its own service identity or on behalf of a user, never with broader permissions than the user who invoked it.\n\n**5. Deterministic workflow state.** The workflow engine, not the model, decides what state a case is in and what happens next. The agent proposes, and the workflow disposes.\n\n**6. Untrusted input.** Content the agent reads (emails, documents, alerts, web pages) is data. Instructions embedded in it are ignored, and attempts are logged.\n\n**7. Full traceability.** Every plan, tool call, input, output, approval and rejection is recorded, so reviewers can reconstruct why something happened.\n\n## Where agents earn their keep\n\n- **Case preparation:** assemble customer, transaction and history context before a human opens the case. See [AI-assisted case management](\u002Fblog\u002Fai-assisted-case-management).\n- **Security enrichment:** read-only lookups across security tools. See [AI in the SOC](\u002Fblog\u002Fai-in-the-soc-triage-and-investigation).\n- **Document workflows:** extract, validate and route documents, and escalate what fails validation.\n- **Operations recovery:** generate and score recovery options for a controller to choose from. See [operations control](\u002Fblog\u002Foperations-control-and-disruption-management).\n- **Reconciliation:** propose matches and classify breaks for an analyst to confirm.\n\nIn each case, the agent compresses the time *before* a human decision. It doesn't replace the decision.\n\n## What to measure\n\nTime saved before the decision point, how often agent proposals are accepted unchanged, the rejection reasons, how often approval gates fire, and incidents caused by agent actions. That last number should be zero, and the design should make it hard to be anything else.\n\n## How it fits the architecture\n\nIn our application foundations, agent tools are ordinary application services with contracts, authorization and tests. That's the same discipline as any other API. This is the practical meaning of [AI accelerates the implementation, architecture governs it](\u002Fblog\u002Fai-accelerates-architecture-governs).\n\n[Bring us a workflow](\u002Fcontact) where an agent could prepare the decision, and we'll scope the checkpoints with you.\n","\u003Cp>Agents, meaning AI systems that plan and take multi-step actions with tools, are the most exciting and the most dangerous AI capability in the enterprise. An agent that gathers context from five systems before an analyst opens a case saves real time. An agent that closes accounts, moves money or emails customers on its own is a governance incident waiting to happen.\u003C\u002Fp>\n\u003Cp>The answer isn't to avoid agents. It's to put them inside a workflow with \u003Cstrong>checkpoints\u003C\u002Fstrong>.\u003C\u002Fp>\n\u003Ch2>Design principles\u003C\u002Fh2>\n\u003Cp>\u003Cstrong>1. Bounded tools.\u003C\u002Fstrong> An agent can only call tools that the application explicitly exposes to it, each with a narrow purpose and validated inputs. No general shell, no arbitrary API access.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>2. Read before write.\u003C\u002Fstrong> Most value comes from read-only work: gathering context, correlating records, drafting. Make read-only the default and treat every write as a separate, higher-risk capability.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>3. Explicit approval for consequential actions.\u003C\u002Fstrong> Anything that changes a record of consequence, contacts a customer, moves value or changes access requires a person to approve. Some actions require two people.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>4. Identity and least privilege.\u003C\u002Fstrong> The agent acts with its own service identity or on behalf of a user, never with broader permissions than the user who invoked it.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>5. Deterministic workflow state.\u003C\u002Fstrong> The workflow engine, not the model, decides what state a case is in and what happens next. The agent proposes, and the workflow disposes.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>6. Untrusted input.\u003C\u002Fstrong> Content the agent reads (emails, documents, alerts, web pages) is data. Instructions embedded in it are ignored, and attempts are logged.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>7. Full traceability.\u003C\u002Fstrong> Every plan, tool call, input, output, approval and rejection is recorded, so reviewers can reconstruct why something happened.\u003C\u002Fp>\n\u003Ch2>Where agents earn their keep\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Cstrong>Case preparation:\u003C\u002Fstrong> assemble customer, transaction and history context before a human opens the case. See \u003Ca href=\"\u002Fblog\u002Fai-assisted-case-management\">AI-assisted case management\u003C\u002Fa>.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Security enrichment:\u003C\u002Fstrong> read-only lookups across security tools. See \u003Ca href=\"\u002Fblog\u002Fai-in-the-soc-triage-and-investigation\">AI in the SOC\u003C\u002Fa>.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Document workflows:\u003C\u002Fstrong> extract, validate and route documents, and escalate what fails validation.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Operations recovery:\u003C\u002Fstrong> generate and score recovery options for a controller to choose from. See \u003Ca href=\"\u002Fblog\u002Foperations-control-and-disruption-management\">operations control\u003C\u002Fa>.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Reconciliation:\u003C\u002Fstrong> propose matches and classify breaks for an analyst to confirm.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>In each case, the agent compresses the time \u003Cem>before\u003C\u002Fem> a human decision. It doesn't replace the decision.\u003C\u002Fp>\n\u003Ch2>What to measure\u003C\u002Fh2>\n\u003Cp>Time saved before the decision point, how often agent proposals are accepted unchanged, the rejection reasons, how often approval gates fire, and incidents caused by agent actions. That last number should be zero, and the design should make it hard to be anything else.\u003C\u002Fp>\n\u003Ch2>How it fits the architecture\u003C\u002Fh2>\n\u003Cp>In our application foundations, agent tools are ordinary application services with contracts, authorization and tests. That's the same discipline as any other API. This is the practical meaning of \u003Ca href=\"\u002Fblog\u002Fai-accelerates-architecture-governs\">AI accelerates the implementation, architecture governs it\u003C\u002Fa>.\u003C\u002Fp>\n\u003Cp>\u003Ca href=\"\u002Fcontact\">Bring us a workflow\u003C\u002Fa> where an agent could prepare the decision, and we'll scope the checkpoints with you.\u003C\u002Fp>\n","Agentic automation with human checkpoints","How to use AI agents in enterprise workflows safely: bounded tools, read-before-write, explicit approvals and an audit trail of every step.",[50,49,62,17],"2026-07-14T00:00:00.000Z",{"id":94,"slug":95,"body":96,"html":97,"title":98,"description":99,"category":11,"tags":100,"author":18,"date":102,"year":20,"month":21,"quarter":22,"status":23,"featured":24},"2026\u002F07\u002Findustry-applications\u002Fcompliance-evidence-produced-by-the-workflow","compliance-evidence-produced-by-the-workflow","\nAsk any compliance team what the week before an audit looks like. Screenshots, exports, email searches and a shared folder that grows until someone declares it complete. The controls probably operated fine. The **evidence** of it was never captured as the work happened.\n\n## The pattern\n\nThe **compliance operations and evidence** family in the Atlas works from a simple principle: every control has an owner, a defined piece of evidence and a system that captures that evidence as a by-product of the work.\n\nA typical foundation includes:\n\n- **Control library.** Controls mapped to obligations, policies and processes, each with an owner and a testing frequency.\n- **Evidence requests and collection.** Scheduled or event-driven, with evidence attached to the control rather than to an email thread.\n- **Attestation workflows.** Owners attest, reviewers challenge and approvers sign off, all with a history.\n- **Exception and issue management.** Failed controls become issues with remediation owners and dates.\n- **Regulatory change intake.** New obligations are assessed and mapped to affected controls.\n- **Reporting and packs.** Audit and supervisory packs generated from the record.\n\n## Where AI helps\n\n- **Document intelligence:** extract the relevant clauses from policies and regulatory texts and propose control mappings for a human to confirm.\n- **Evidence classification:** check that an uploaded file actually matches what the control requires, and flag mismatches before a reviewer finds them.\n- **Summarization:** turn a quarter of attestations and issues into a readable management summary.\n- **Gap detection:** highlight controls with stale or missing evidence ahead of the audit.\n\nThe application records who accepted or rejected every AI suggestion. The AI never attests.\n\n## Who uses it\n\nCompliance officers, control owners across the business, internal audit, risk officers and, in the public sector, inspection and oversight teams.\n\n## Integrations\n\nTicketing and ITSM, where much evidence already lives. Document management. The identity provider, so attestations are tied to real people. HR systems for ownership changes. Data platforms for automated control tests.\n\n## The difference it makes\n\nAn evidence application changes the question from “can we prove it?” to “show me the record.” It also changes the economics. The effort moves from assembling evidence to operating controls, which is where it should have been all along.\n\n## Where it applies\n\nBanking and insurance, payments, government entities with internal-control obligations, and any organization with recurring audits (ISO, SOC or sector regulators). For licensed digital-asset operators, the same foundation handles KYC, KYT and Travel Rule operations. See [digital assets](\u002Findustries\u002Fdigital-assets).\n\n## A sensible first scope\n\nOne control domain, such as access reviews or third-party oversight, with its evidence moved into the application ahead of the next audit cycle. Scope it in a [Solution Definition Sprint](\u002Fservices\u002Fsolution-definition-sprint), or [bring us the audit you dread most](\u002Fcontact).\n\n*fazeZERO builds and integrates applications. Regulatory interpretation stays with your compliance function and counsel.*\n","\u003Cp>Ask any compliance team what the week before an audit looks like. Screenshots, exports, email searches and a shared folder that grows until someone declares it complete. The controls probably operated fine. The \u003Cstrong>evidence\u003C\u002Fstrong> of it was never captured as the work happened.\u003C\u002Fp>\n\u003Ch2>The pattern\u003C\u002Fh2>\n\u003Cp>The \u003Cstrong>compliance operations and evidence\u003C\u002Fstrong> family in the Atlas works from a simple principle: every control has an owner, a defined piece of evidence and a system that captures that evidence as a by-product of the work.\u003C\u002Fp>\n\u003Cp>A typical foundation includes:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>Control library.\u003C\u002Fstrong> Controls mapped to obligations, policies and processes, each with an owner and a testing frequency.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Evidence requests and collection.\u003C\u002Fstrong> Scheduled or event-driven, with evidence attached to the control rather than to an email thread.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Attestation workflows.\u003C\u002Fstrong> Owners attest, reviewers challenge and approvers sign off, all with a history.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Exception and issue management.\u003C\u002Fstrong> Failed controls become issues with remediation owners and dates.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Regulatory change intake.\u003C\u002Fstrong> New obligations are assessed and mapped to affected controls.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Reporting and packs.\u003C\u002Fstrong> Audit and supervisory packs generated from the record.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>Where AI helps\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Cstrong>Document intelligence:\u003C\u002Fstrong> extract the relevant clauses from policies and regulatory texts and propose control mappings for a human to confirm.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Evidence classification:\u003C\u002Fstrong> check that an uploaded file actually matches what the control requires, and flag mismatches before a reviewer finds them.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Summarization:\u003C\u002Fstrong> turn a quarter of attestations and issues into a readable management summary.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Gap detection:\u003C\u002Fstrong> highlight controls with stale or missing evidence ahead of the audit.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>The application records who accepted or rejected every AI suggestion. The AI never attests.\u003C\u002Fp>\n\u003Ch2>Who uses it\u003C\u002Fh2>\n\u003Cp>Compliance officers, control owners across the business, internal audit, risk officers and, in the public sector, inspection and oversight teams.\u003C\u002Fp>\n\u003Ch2>Integrations\u003C\u002Fh2>\n\u003Cp>Ticketing and ITSM, where much evidence already lives. Document management. The identity provider, so attestations are tied to real people. HR systems for ownership changes. Data platforms for automated control tests.\u003C\u002Fp>\n\u003Ch2>The difference it makes\u003C\u002Fh2>\n\u003Cp>An evidence application changes the question from “can we prove it?” to “show me the record.” It also changes the economics. The effort moves from assembling evidence to operating controls, which is where it should have been all along.\u003C\u002Fp>\n\u003Ch2>Where it applies\u003C\u002Fh2>\n\u003Cp>Banking and insurance, payments, government entities with internal-control obligations, and any organization with recurring audits (ISO, SOC or sector regulators). For licensed digital-asset operators, the same foundation handles KYC, KYT and Travel Rule operations. See \u003Ca href=\"\u002Findustries\u002Fdigital-assets\">digital assets\u003C\u002Fa>.\u003C\u002Fp>\n\u003Ch2>A sensible first scope\u003C\u002Fh2>\n\u003Cp>One control domain, such as access reviews or third-party oversight, with its evidence moved into the application ahead of the next audit cycle. Scope it in a \u003Ca href=\"\u002Fservices\u002Fsolution-definition-sprint\">Solution Definition Sprint\u003C\u002Fa>, or \u003Ca href=\"\u002Fcontact\">bring us the audit you dread most\u003C\u002Fa>.\u003C\u002Fp>\n\u003Cp>\u003Cem>fazeZERO builds and integrates applications. Regulatory interpretation stays with your compliance function and counsel.\u003C\u002Fem>\u003C\u002Fp>\n","Compliance evidence should be produced by the workflow, not assembled for the audit","Regulatory evidence collection and control attestation as an application: controls mapped to evidence, captured as work happens, reviewed by owners.",[101,17,14,82,16],"compliance","2026-07-09T00:00:00.000Z",{"id":104,"slug":105,"body":106,"html":107,"title":108,"description":109,"category":59,"tags":110,"author":18,"date":112,"year":20,"month":21,"quarter":22,"status":23,"featured":24,"series":113,"seriesOrder":114},"2026\u002F07\u002Fai-in-production\u002Fdeployment-ready-is-not-production","deployment-ready-is-not-production","\n“Production-ready” is one of the most abused phrases in enterprise software. It usually means “the demo worked.” We avoid it on purpose.\n\nWe describe every application in one of three stages. Each stage describes what has actually happened, not what we hope will happen.\n\n## 1. Deployment-Ready Application Foundation\n\nThis is where every application in our inventory starts. A foundation has:\n\n- been generated and scaffolded on the factory architecture\n- a runnable core with domain services and adapters\n- an API structure defined by contracts\n- a web application\n- identity and authorization patterns\n- automated tests\n- a deployment baseline\n- a known, consistent architectural structure\n\nA foundation is real software, not a mock-up. But it hasn't met your data, your identity provider, your integrations or your policies. Calling it “production” would be misleading.\n\n## 2. Customer-Configured Application\n\nThe foundation has been adapted to one organization:\n\n- the customer's requirements and business workflows\n- their data and data platforms\n- integrations with their systems of record\n- AI and model choices, including providers, hosting and evaluation criteria\n- their identity environment\n- their cloud platform and regional constraints\n- their policies and approval rules\n- their operating environment\n\nThis is what an [AI Production Sprint](\u002Fservices\u002Fai-production-sprint) delivers. It is a working application on a production path, but it isn't in production yet.\n\n## 3. Production Deployment\n\nThe application has completed the customer-specific work that production actually requires:\n\n- integration testing against live systems\n- security hardening and review\n- cloud deployment in the customer's environment\n- operational testing\n- customer acceptance\n- observability and alerting\n- a support design\n- production controls\n\nOnly then do we call it production.\n\n## Why the distinction matters\n\n**For buyers**, it sets honest expectations. A foundation shortens the path to production. It doesn't remove the path. Integration, hardening and acceptance still take real effort, and a vendor who claims otherwise is either underestimating your environment or overestimating their template.\n\n**For risk and security teams**, it gives a shared vocabulary. A deployment-ready foundation can be reviewed architecturally. A customer-configured application can be reviewed against your policies. A production deployment has passed your gates.\n\n**For partners**, it prevents over-selling. A systems integrator who promises a “production-ready AI app in two weeks” is setting up the relationship to fail.\n\n## How the stages map to engagements\n\n| Stage | How you get there |\n|---|---|\n| Deployment-ready foundation | Already in the inventory, or generated by the factory |\n| Customer-configured | [Solution Definition Sprint](\u002Fservices\u002Fsolution-definition-sprint), then [AI Production Sprint](\u002Fservices\u002Fai-production-sprint) |\n| Production deployment | Completion of the production roadmap, often inside an [Application Family Program](\u002Fservices\u002Fapplication-family-program) |\n\n## What we won't do\n\nWe won't label a listing “production” because it runs. We won't describe a customer-configured application as a production deployment before acceptance. And we won't publish customer names, metrics or badges we can't evidence.\n\nIt's a small discipline. It also happens to be the one enterprise buyers trust most.\n\nRead more about [the factory](\u002Ffactory), or [bring us a use case](\u002Fcontact).\n","\u003Cp>“Production-ready” is one of the most abused phrases in enterprise software. It usually means “the demo worked.” We avoid it on purpose.\u003C\u002Fp>\n\u003Cp>We describe every application in one of three stages. Each stage describes what has actually happened, not what we hope will happen.\u003C\u002Fp>\n\u003Ch2>1. Deployment-Ready Application Foundation\u003C\u002Fh2>\n\u003Cp>This is where every application in our inventory starts. A foundation has:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>been generated and scaffolded on the factory architecture\u003C\u002Fli>\n\u003Cli>a runnable core with domain services and adapters\u003C\u002Fli>\n\u003Cli>an API structure defined by contracts\u003C\u002Fli>\n\u003Cli>a web application\u003C\u002Fli>\n\u003Cli>identity and authorization patterns\u003C\u002Fli>\n\u003Cli>automated tests\u003C\u002Fli>\n\u003Cli>a deployment baseline\u003C\u002Fli>\n\u003Cli>a known, consistent architectural structure\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>A foundation is real software, not a mock-up. But it hasn't met your data, your identity provider, your integrations or your policies. Calling it “production” would be misleading.\u003C\u002Fp>\n\u003Ch2>2. Customer-Configured Application\u003C\u002Fh2>\n\u003Cp>The foundation has been adapted to one organization:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>the customer's requirements and business workflows\u003C\u002Fli>\n\u003Cli>their data and data platforms\u003C\u002Fli>\n\u003Cli>integrations with their systems of record\u003C\u002Fli>\n\u003Cli>AI and model choices, including providers, hosting and evaluation criteria\u003C\u002Fli>\n\u003Cli>their identity environment\u003C\u002Fli>\n\u003Cli>their cloud platform and regional constraints\u003C\u002Fli>\n\u003Cli>their policies and approval rules\u003C\u002Fli>\n\u003Cli>their operating environment\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>This is what an \u003Ca href=\"\u002Fservices\u002Fai-production-sprint\">AI Production Sprint\u003C\u002Fa> delivers. It is a working application on a production path, but it isn't in production yet.\u003C\u002Fp>\n\u003Ch2>3. Production Deployment\u003C\u002Fh2>\n\u003Cp>The application has completed the customer-specific work that production actually requires:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>integration testing against live systems\u003C\u002Fli>\n\u003Cli>security hardening and review\u003C\u002Fli>\n\u003Cli>cloud deployment in the customer's environment\u003C\u002Fli>\n\u003Cli>operational testing\u003C\u002Fli>\n\u003Cli>customer acceptance\u003C\u002Fli>\n\u003Cli>observability and alerting\u003C\u002Fli>\n\u003Cli>a support design\u003C\u002Fli>\n\u003Cli>production controls\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>Only then do we call it production.\u003C\u002Fp>\n\u003Ch2>Why the distinction matters\u003C\u002Fh2>\n\u003Cp>\u003Cstrong>For buyers\u003C\u002Fstrong>, it sets honest expectations. A foundation shortens the path to production. It doesn't remove the path. Integration, hardening and acceptance still take real effort, and a vendor who claims otherwise is either underestimating your environment or overestimating their template.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>For risk and security teams\u003C\u002Fstrong>, it gives a shared vocabulary. A deployment-ready foundation can be reviewed architecturally. A customer-configured application can be reviewed against your policies. A production deployment has passed your gates.\u003C\u002Fp>\n\u003Cp>\u003Cstrong>For partners\u003C\u002Fstrong>, it prevents over-selling. A systems integrator who promises a “production-ready AI app in two weeks” is setting up the relationship to fail.\u003C\u002Fp>\n\u003Ch2>How the stages map to engagements\u003C\u002Fh2>\n\u003Ctable>\n\u003Cthead>\n\u003Ctr>\n\u003Cth>Stage\u003C\u002Fth>\n\u003Cth>How you get there\u003C\u002Fth>\n\u003C\u002Ftr>\n\u003C\u002Fthead>\n\u003Ctbody>\u003Ctr>\n\u003Ctd>Deployment-ready foundation\u003C\u002Ftd>\n\u003Ctd>Already in the inventory, or generated by the factory\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Customer-configured\u003C\u002Ftd>\n\u003Ctd>\u003Ca href=\"\u002Fservices\u002Fsolution-definition-sprint\">Solution Definition Sprint\u003C\u002Fa>, then \u003Ca href=\"\u002Fservices\u002Fai-production-sprint\">AI Production Sprint\u003C\u002Fa>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003Ctr>\n\u003Ctd>Production deployment\u003C\u002Ftd>\n\u003Ctd>Completion of the production roadmap, often inside an \u003Ca href=\"\u002Fservices\u002Fapplication-family-program\">Application Family Program\u003C\u002Fa>\u003C\u002Ftd>\n\u003C\u002Ftr>\n\u003C\u002Ftbody>\u003C\u002Ftable>\n\u003Ch2>What we won't do\u003C\u002Fh2>\n\u003Cp>We won't label a listing “production” because it runs. We won't describe a customer-configured application as a production deployment before acceptance. And we won't publish customer names, metrics or badges we can't evidence.\u003C\u002Fp>\n\u003Cp>It's a small discipline. It also happens to be the one enterprise buyers trust most.\u003C\u002Fp>\n\u003Cp>Read more about \u003Ca href=\"\u002Ffactory\">the factory\u003C\u002Fa>, or \u003Ca href=\"\u002Fcontact\">bring us a use case\u003C\u002Fa>.\u003C\u002Fp>\n","Deployment-ready is not production: an honest maturity model","Why we describe applications in three stages (deployment-ready foundation, customer-configured, production deployment) and never call a template 'production'.",[62,36,111,16],"implementation","2026-07-07T00:00:00.000Z","the-application-factory",5,{"id":116,"slug":117,"body":118,"html":119,"title":120,"description":121,"category":11,"tags":122,"author":18,"date":125,"year":20,"month":21,"quarter":22,"status":23,"featured":24},"2026\u002F07\u002Findustry-applications\u002Fai-model-governance-as-an-application","ai-model-governance-as-an-application","\nMost enterprises now have an AI policy. Far fewer have an AI governance **system**. The policy says every model must be inventoried, evaluated, approved and monitored. In practice, the inventory is a spreadsheet, the evaluations are in notebooks, approvals happen in email and monitoring depends on whoever built the model.\n\nThat works for five models. It fails at fifty, and it fails immediately when an auditor or supervisor asks for evidence.\n\n## The workflow behind “AI governance”\n\nThe **AI governance** family in the Atlas treats governance as an operational workflow with a system of record:\n\n1. **Register.** Every model and AI use case gets an owner, a purpose, a risk tier, its data sources and where it is deployed. That includes vendor models, LLM features and internal models.\n2. **Evaluate.** Structured evaluations against defined criteria: accuracy, robustness, bias and fairness, and for LLM features, groundedness and safety. Results are stored as evidence, not screenshots.\n3. **Approve.** Deployment requests route through the right reviewers, such as model risk, security, the business owner and compliance, based on the risk tier. Every decision is recorded.\n4. **Monitor.** Production behaviour is tracked against thresholds. Drift and incidents raise cases with owners.\n5. **Evidence.** Packs for internal audit, the board or supervisors are generated from the record.\n\n## Where AI helps inside the governance application\n\nIt sounds recursive, but it's useful:\n\n- **Summarization** of model documentation and evaluation results for reviewers\n- **Classification** of new use cases into risk tiers, as a suggestion for a human to confirm\n- **Evaluation assistance**, generating test cases and red-team prompts for LLM features\n- **Drafting** evidence-pack narratives from structured records\n\nEvery one of these is a draft for a human. The approval decision is never automated.\n\n## Who uses it\n\n- **Head of AI and the AI platform team:** keep the portfolio visible and deployable.\n- **Model risk managers:** run reviews with consistent criteria.\n- **Risk and compliance officers:** answer supervisors and auditors from one record.\n- **CIO, CDO and CDAO:** see where AI is used, by whom, and at what risk.\n\n## Integrations that matter\n\nModel registries and ML platforms, CI\u002FCD pipelines (so deployment approval is a real gate rather than a formality), the identity provider for reviewer roles, ticketing, and data catalogues for lineage.\n\n## Controls designed in\n\n- Segregation between model owner and approver\n- An immutable decision history\n- Required evidence before approval can proceed\n- Periodic re-review based on risk tier and staleness\n- Role-based access to sensitive evaluation data\n\n## Why it belongs in financial services first\n\nBanks and insurers already run model risk management for credit and pricing models. Generative AI has multiplied the number of “models” and blurred their edges. A governance application extends existing discipline to the new portfolio instead of creating a parallel process.\n\nThe same foundation applies across enterprise operations, government and any organization preparing for AI-specific regulation.\n\n## Starting point\n\nThe fastest start is to take one line of business's AI inventory and move it into the application, with the approval workflow switched on for new deployments only. The [Solution Definition Sprint](\u002Fservices\u002Fsolution-definition-sprint) scopes the delta: your risk tiers, reviewers, evaluation criteria and integrations.\n\nSee the [financial services](\u002Findustries\u002Ffinancial-services) page, search the [Atlas](\u002Fatlas), or [bring us your AI inventory](\u002Fcontact).\n","\u003Cp>Most enterprises now have an AI policy. Far fewer have an AI governance \u003Cstrong>system\u003C\u002Fstrong>. The policy says every model must be inventoried, evaluated, approved and monitored. In practice, the inventory is a spreadsheet, the evaluations are in notebooks, approvals happen in email and monitoring depends on whoever built the model.\u003C\u002Fp>\n\u003Cp>That works for five models. It fails at fifty, and it fails immediately when an auditor or supervisor asks for evidence.\u003C\u002Fp>\n\u003Ch2>The workflow behind “AI governance”\u003C\u002Fh2>\n\u003Cp>The \u003Cstrong>AI governance\u003C\u002Fstrong> family in the Atlas treats governance as an operational workflow with a system of record:\u003C\u002Fp>\n\u003Col>\n\u003Cli>\u003Cstrong>Register.\u003C\u002Fstrong> Every model and AI use case gets an owner, a purpose, a risk tier, its data sources and where it is deployed. That includes vendor models, LLM features and internal models.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Evaluate.\u003C\u002Fstrong> Structured evaluations against defined criteria: accuracy, robustness, bias and fairness, and for LLM features, groundedness and safety. Results are stored as evidence, not screenshots.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Approve.\u003C\u002Fstrong> Deployment requests route through the right reviewers, such as model risk, security, the business owner and compliance, based on the risk tier. Every decision is recorded.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Monitor.\u003C\u002Fstrong> Production behaviour is tracked against thresholds. Drift and incidents raise cases with owners.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Evidence.\u003C\u002Fstrong> Packs for internal audit, the board or supervisors are generated from the record.\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Ch2>Where AI helps inside the governance application\u003C\u002Fh2>\n\u003Cp>It sounds recursive, but it's useful:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>Summarization\u003C\u002Fstrong> of model documentation and evaluation results for reviewers\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Classification\u003C\u002Fstrong> of new use cases into risk tiers, as a suggestion for a human to confirm\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Evaluation assistance\u003C\u002Fstrong>, generating test cases and red-team prompts for LLM features\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Drafting\u003C\u002Fstrong> evidence-pack narratives from structured records\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>Every one of these is a draft for a human. The approval decision is never automated.\u003C\u002Fp>\n\u003Ch2>Who uses it\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Cstrong>Head of AI and the AI platform team:\u003C\u002Fstrong> keep the portfolio visible and deployable.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Model risk managers:\u003C\u002Fstrong> run reviews with consistent criteria.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Risk and compliance officers:\u003C\u002Fstrong> answer supervisors and auditors from one record.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>CIO, CDO and CDAO:\u003C\u002Fstrong> see where AI is used, by whom, and at what risk.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>Integrations that matter\u003C\u002Fh2>\n\u003Cp>Model registries and ML platforms, CI\u002FCD pipelines (so deployment approval is a real gate rather than a formality), the identity provider for reviewer roles, ticketing, and data catalogues for lineage.\u003C\u002Fp>\n\u003Ch2>Controls designed in\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>Segregation between model owner and approver\u003C\u002Fli>\n\u003Cli>An immutable decision history\u003C\u002Fli>\n\u003Cli>Required evidence before approval can proceed\u003C\u002Fli>\n\u003Cli>Periodic re-review based on risk tier and staleness\u003C\u002Fli>\n\u003Cli>Role-based access to sensitive evaluation data\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2>Why it belongs in financial services first\u003C\u002Fh2>\n\u003Cp>Banks and insurers already run model risk management for credit and pricing models. Generative AI has multiplied the number of “models” and blurred their edges. A governance application extends existing discipline to the new portfolio instead of creating a parallel process.\u003C\u002Fp>\n\u003Cp>The same foundation applies across enterprise operations, government and any organization preparing for AI-specific regulation.\u003C\u002Fp>\n\u003Ch2>Starting point\u003C\u002Fh2>\n\u003Cp>The fastest start is to take one line of business's AI inventory and move it into the application, with the approval workflow switched on for new deployments only. The \u003Ca href=\"\u002Fservices\u002Fsolution-definition-sprint\">Solution Definition Sprint\u003C\u002Fa> scopes the delta: your risk tiers, reviewers, evaluation criteria and integrations.\u003C\u002Fp>\n\u003Cp>See the \u003Ca href=\"\u002Findustries\u002Ffinancial-services\">financial services\u003C\u002Fa> page, search the \u003Ca href=\"\u002Fatlas\">Atlas\u003C\u002Fa>, or \u003Ca href=\"\u002Fcontact\">bring us your AI inventory\u003C\u002Fa>.\u003C\u002Fp>\n","AI model governance should be an application, not a policy document","Model inventory, evaluation, deployment approval and monitoring as one governed workflow, so AI governance produces evidence instead of meetings.",[123,14,124,17,13],"ai-governance","evaluation","2026-07-02T00:00:00.000Z",1790080514008]