{"resourceId":"route-fifty-urban-agentic-ai-playbook-20260917","versions":[{"version":"external-07ca7887016c823d3511a8b3da4af7f14eeefbc68d4844e328f669b3f9499653","resource":{"id":"route-fifty-urban-agentic-ai-playbook-20260917","title":"Route Fifty: Urban Institute Responsible Agentic AI Playbook urges roles, audit trails, internal vs public-facing gates","organization":"Urban Institute (Graham MacDonald) as covered by Route Fifty","sector":"Independent public-policy research / state-local AI governance guidance","geography":"U.S. state and local governments (examples: Virginia regulatory agents, Utah Rx renewal pilot, Boston open data)","publishedAt":"2026-09-17","publicationDate":"2026-09-17","eventDate":"2026-09-17","sourceName":"Route Fifty","sourceLabel":"Urban Institute releases guidance for state and local agentic AI adoption (Kaitlyn Levinson)","sourceUrl":"https://www.route-fifty.com/artificial-intelligence/2026/09/urban-institute-releases-guidance-state-and-local-agentic-ai-adoption/416054/","evidenceClass":"independent-reporting","outcomeClass":"emerging","topics":["developers-agents","governance-procurement","operating-model","accessibility-workforce"],"finding":"Route Fifty (2026-09-17) covers Urban Institute CIO Graham MacDonald's Responsible Agentic AI Playbook for state/local builders. Agentic systems are described as having autonomy to plan and take action beyond content generation. The playbook distinguishes internal-facing vs constituent-facing agents (stricter fairness/bias testing for public use), urges documentation of purpose/population/outcomes/costs, and names seven governance roles (accountable owner, evaluation lead, security lead, transparency lead, responsible AI lead, contributor, independent review) with tasks such as maintaining audit trails and monitoring harms. Reporting cites Virginia's regulatory-document agentic pilot (claimed more than $1.4B annual savings / about 80% permit-time cut per a January ORM report) and Utah's Doctronic Rx-renewal pilot (72% suggested renewals; physicians agreed with 91% of those) as early examples—not as audited Route Fifty findings.","sledRelevance":"Independent governance guidance timed against NASCIO's first agentic-AI survey questions—gives state buyers concrete role and audit-trail expectations. Why now: not archived; newly relevant the week NASCIO reports 64% of CIOs expect agentic AI to be most impactful.","evidence":"Opened the Route Fifty HTTPS article via WebFetch/curl; confirmed September 17, 2026; captured MacDonald definition of agentic autonomy, internal vs external gates, seven roles, audit-trail tasks, and Virginia/Utah example figures as attributed in the article. Also retrieved the Urban GitHub playbook markdown for corroboration of non-expert/internal-external framing.","architectureImplications":"Design agent stacks with explicit tool permissions, evaluation harnesses, and separate promotion criteria for internal vs public agents; retain audit trails of plans/actions.","governanceImplications":"Assign the seven roles (or map them onto existing AI committees) before production; require purpose/metrics artifacts at design time, not after incidents.","securityPrivacyImplications":"Security lead and least-privilege tool access are first-class; public-facing agents need stronger fairness testing because autonomy can change resident outcomes.","caveats":"Journalism summarizing a living playbook; Virginia $1.4B / 80% and Utah 72%/91% figures are attributed to prior state/vendor reports, not independently verified here; playbook is guidance, not a statute or measured multi-state evaluation.","streamIds":["state-government","local-government"],"roles":{"sales":"Frame Urban's agentic playbook—via Route Fifty—as a governance readiness conversation for states eyeing autonomy beyond chat, not as endorsement of Virginia's savings headline. Discovery should ask whether the use is internal or constituent-facing and which of the seven roles already exist. Bounded engagement: map one candidate agent to purpose/metrics artifacts and an evaluation rubric before build. Unsupported claims include transferring Virginia's $1.4B figure, that seven roles guarantee safety, or that internal agents need no fairness testing when outputs later inform public decisions.","engineering":"Fit is an agent architecture with tool permissions, eval harness, and promotion gates that tighten for public-facing use. Architecture should log plans/actions, separate internal sandboxes from constituent channels, and define accuracy/fairness metrics per MacDonald's framing. Prerequisites: accountable owner, evaluation lead, security review. Constraints: playbook is aspirational; example metrics are second-hand. Proof of value is passing a documented eval rubric with audit-trail completeness—not demo autonomy.","delivery":"Delivery establishes role assignments, writes purpose/population/outcome docs, and runs internal pilot before any public agent. Dependencies: CIO/data science, program owner, security, independent review capacity. Ownership of harms monitoring sits with the accountable owner. Acceptance criteria: role roster published, audit trails queryable, fairness tests for public agents. Risks include human-in-loop theater when reviewers lack system literacy, and importing vendor autonomy without eval templates."},"roleDetails":{"sales":{"customerProblem":"States want agentic efficiency but lack a practical standard for roles, evals, and internal vs public gates.","stakeholders":["State/local CIO","Program owners","Security/privacy","Evaluation/QA","Independent review / auditors"],"discoveryQuestions":["Is the agent internal-only or constituent-facing?","Who is the accountable owner and evaluation lead?","What fairness/accuracy metrics are defined before build?","Can you export an audit trail of agent actions today?"],"valueHypothesis":"Documented roles and eval rubrics reduce agentic harm risk enough to justify a bounded internal pilot before public scale.","potentialEngagement":"Playbook gap assessment and one internal agent design package with eval criteria.","unsupportedClaims":["Virginia $1.4B savings transferable","Seven roles guarantee safe autonomy","Internal agents never need fairness tests","Playbook supersedes state law"]},"engineering":{"fit":"Tool-calling agents with permissioning, eval harness, and stricter public-facing promotion gates.","architectureIntegration":"Planner/tools to permission broker to action log to human review for high-impact steps to separate public endpoint.","prerequisites":["Accountable owner","Eval rubric","Security lead sign-off","Corpus/data access rules"],"constraints":["Guidance not statute","Example KPIs second-hand","Reviewer literacy gaps"],"security":"Least-privilege tools, secrets isolation, continuous monitoring for harmful actions.","proofOfValue":["Eval rubric pass rate","Audit-trail completeness sample","Fairness metric for public agent if applicable"]},"delivery":{"work":"Assign seven roles (or map equivalents); draft purpose/metrics; run internal pilot; decide public promotion.","dependencies":["Program owner","Security","Eval capacity","Independent review"],"ownership":"Accountable owner owns production agent outcomes and harm monitoring.","skillsAdoption":"Train human reviewers on how the agent plans/acts so oversight is informed.","governanceCheckpoints":["Design artifact review","Internal pilot eval","Public-facing fairness gate","Ongoing harm monitoring"],"acceptanceCriteria":["Role roster published","Queryably logged actions","Documented go/no-go for public use"],"risks":["HITL without reviewer literacy","Autonomy without eval templates","Over-claiming example state savings"]}},"retrievedAt":"2026-09-29T22:20:24-05:00","enrichedAt":"2026-09-29T22:20:24-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Bringing non-IT staff into agent building expands access but raises skill gaps for human reviewers who must understand the system they oversee.","procurementImplications":"RFPs for agentic tools should require role assignments, evaluation rubrics, audit-trail export, and fairness-test evidence for public-facing agents.","operatingModelImplications":"Start with smaller-scale internal agents; escalate governance rigor when agents face constituents or benefits/health workflows.","sourceVerification":{"responseId":"webfetch-route50-urban-agentic-20260929T222024","openedUrl":"https://www.route-fifty.com/artificial-intelligence/2026/09/urban-institute-releases-guidance-state-and-local-agentic-ai-adoption/416054/","referenceExcerpt":"maintaining audit trails of an agentic AI system, assessing the tool against evaluation templates or rubrics","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}