{"resourceId":"maryland-apep-nascio-dhs-2026","versions":[{"version":"external-7c590a9478e1a5ae0a6bde8c9bc5125c01cacbda7e71cc4d7555dfc7ac2592a2","resource":{"id":"maryland-apep-nascio-dhs-2026","title":"Maryland DHS APEP NASCIO: grounded policy GenAI reports 88.9% citation accuracy and statewide Spring 2026 launch","organization":"Maryland Department of Human Services (DHS) / Office of Technology for Human Services","sector":"State human services / caseworker policy enablement / generative AI training tools","geography":"Maryland statewide DHS (Alpha pilot Family Investment; confirmatory pilot Baltimore City; Spring 2026 statewide)","publishedAt":"2026-08-01","publicationDate":"2026-08-01","eventDate":"2026-04-01","sourceName":"NASCIO / Maryland DHS","sourceLabel":"AI-Powered Policy Enablement Package (APEP)","sourceUrl":"https://www.nascio.org/wp-content/uploads/2026/08/MD_ArtificialIntelligence.pdf","evidenceClass":"public-sector-association","outcomeClass":"effective","topics":["knowledge-work","accessibility-workforce","developers-agents","operating-model"],"finding":"Maryland DHS’s NASCIO AI nomination for APEP describes a multi-modal policy enablement suite (synthesized audio/video, visual artifacts, grounded Policy Bot, Adaptive Policy Simulator) moving from Dec 2025 pilot to Spring 2026 statewide launch. Success metrics target >80% correct citations, <60 minutes per package, and >75% user satisfaction. Reported results: 88.9% technical accuracy in policy guidance during initial testing; training-material development cut from weeks to <30 minutes for experienced policy experts; 77.8% usefulness score; 100% savings on new software by using existing enterprise licenses. Human experts review AI content; caseworkers—not AI—make eligibility decisions; out-of-bounds testing refuses unsupported questions.","sledRelevance":"Held over from 2026-09-30 award-week research as the strongest unarchived measured policy-bot nomination. Why now: still not in archive after GovTech awards coverage; adds accuracy/usefulness metrics for human-services GenAI that complement Ohio/NYS award entries.","evidence":"Opened NASCIO HTTPS PDF via curl + pdftotext; confirmed pilot timeline, four-part toolset, HITL/augmentation claims, Payment Accuracy Program embedding, and Impact metrics (88.9%, <30 minutes, 77.8%, license reuse).","architectureImplications":"Grounded RAG-style Policy Bot plus simulator sandbox on private departmental cloud patterns; keep eligibility determination outside the model; log refusals for out-of-bounds queries.","governanceImplications":"Decentralized authoring by policy experts with mandatory human review shifts IT from builder to platform governor—inventory and accuracy sampling remain necessary.","securityPrivacyImplications":"Human-services policy content is sensitive; require private grounding corpora, no resident case data in training content generation without authorization, and access controls for simulator scenarios.","caveats":"Self-reported nomination; 88.9% is initial testing accuracy against a >80% target—not a published RCT; usefulness 77.8% is pilot-scoped; eligibility outcomes not claimed as AI-decided.","streamIds":["state-government"],"roles":{"sales":"Frame APEP as a measured human-services policy GenAI pattern—88.9% citation accuracy in testing and Spring 2026 statewide launch with HITL—not as AI eligibility adjudication. Discovery should ask how accuracy is sampled, whether case data is excluded from grounding, and who owns policy-author review. Bounded engagement: a 60–90 day grounded policy-assistant pilot on one program area with citation audit and usefulness survey. Unsupported claims include guaranteed 88.9% production accuracy, AI deciding benefits, or that license reuse always equals zero cost.","engineering":"Fit is a grounded multi-modal policy assistant and simulator for caseworker enablement with human review—not an eligibility decision engine. Architecture: private corpus RAG/policy bot, content-generation pipeline with reviewer workflow, simulator sandbox, inventory logging. Prerequisites: authoritative policy corpus, reviewer roster, refusal/out-of-bounds tests, enterprise licenses. Constraints: nomination self-report; initial-test accuracy only. Proof of value is ongoing citation audit ≥target and measured time-to-package, with eligibility decisions remaining human.","delivery":"Delivery stands up authoring/review roles, migrates priority policy packages, runs accuracy sampling, and ties outputs to Payment Accuracy / training ops. Dependencies: DHS policy, OTHS, county/local offices, privacy, procurement. Ownership of published policy guidance stays with human policy experts. Acceptance criteria: documented accuracy sample, reviewer sign-off path, and no AI eligibility determinations. Risks: accuracy drift after statewide scale and mistaking training speed for correct field decisions."},"roleDetails":{"sales":{"customerProblem":"Human-services agencies struggle to close the readiness gap when policy changes faster than caseworker training.","stakeholders":["DHS/HHS leadership","Policy authors","Casework supervisors","CIO/OTHS","Quality/Payment Accuracy"],"discoveryQuestions":["How is citation accuracy sampled in production?","Is resident case data excluded from grounding?","Who approves AI-generated training packages?","What is the refusal behavior for out-of-scope questions?"],"valueHypothesis":"Grounded policy GenAI with HITL can raise citation accuracy and cut package build time without automating eligibility decisions.","potentialEngagement":"60–90 day policy-bot pilot on one benefit program with citation audit.","unsupportedClaims":["Guaranteed 88.9% production accuracy","AI replaces caseworker judgment","Zero ongoing license/ops cost","Statewide launch equals outcome evaluation complete"]},"engineering":{"fit":"Grounded policy assistant + simulator with mandatory human review; no eligibility automation.","architectureIntegration":"Private corpus to GenAI package pipeline to reviewer workflow to caseworker tools to inventory logs.","prerequisites":["Authoritative policy corpus","Reviewer roster","Out-of-bounds tests","Enterprise GenAI licenses"],"constraints":["Self-reported nomination","Initial-test accuracy only","Usefulness pilot-scoped"],"security":"Private tenancy; minimize case PII in content gen; access control on simulator scenarios.","proofOfValue":["Citation audit vs gold answers","Time-to-package measurement","Usefulness survey ≥target"]},"delivery":{"work":"Select program; author packages; sample accuracy; train supervisors; monitor refusals.","dependencies":["Policy office","OTHS","Field offices","Privacy","Training ops"],"ownership":"Policy experts own published guidance; IT owns platform.","skillsAdoption":"Convert authors to reviewers; coach caseworkers on verify-then-act.","governanceCheckpoints":["Corpus freeze","Accuracy gate","Statewide expand review"],"acceptanceCriteria":["Accuracy sample documented","Reviewer sign-off recorded","No AI eligibility decisions"],"risks":["Accuracy drift","Field over-trust","Content sprawl without inventory"]}},"retrievedAt":"2026-10-01T22:23:45-05:00","enrichedAt":"2026-10-01T22:23:45-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Audio/video and simulator modes support mobile caseworker learning; usefulness metric suggests adoption still needs coaching for emotional-intelligence practice, not only citation lookup.","procurementImplications":"Claimed 100% new-software savings via existing licenses implies prioritize enterprise GenAI/platform reuse and SOW clarity on grounding/accuracy SLAs over net-new point tools.","operatingModelImplications":"Treat policy enablement AI as a line-of-business capability with reviewer roles and Payment Accuracy linkage—not a one-time IT pilot.","sourceVerification":{"responseId":"curl-md-apep-20261001T222345","openedUrl":"https://www.nascio.org/wp-content/uploads/2026/08/MD_ArtificialIntelligence.pdf","referenceExcerpt":"Verified 88.9% accuracy in policy guidance during initial testing.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}