Lighthouse AdvisorySLED AI Adoption Intelligence

State & Local GovernmentIssue 204 records

State Government

One day after Digital States 2026 and Utah’s Gemini hours-saved interview, three more A-grade state CIO Q&As (2026-09-23) show GenAI moving from frameworks into measured operations—and a same-day legislative tracker keeps oversight pressure visible. Michigan Acting CIO Eric Swanson reports MDHHS food-assistance document AI at up to 97% (digital) / 73% (handwritten) automation and ~266 staff hours/week reduced, with eligibility still human-reviewed, plus DTMB coding-assistant productivity claims; Nevada CIO Timothy Galluzi highlights enterprise Copilot plus AI call-center and cyber tooling; Georgia CIO Shawnzia Thomas stresses Office of AI / Horizons Innovation Lab governance-first scale. Transparency Coalition’s 2026-09-25 weekly update tallies 85 new AI-related laws in 27 states so far in 2026 and notes California AB 1609 (customer-service chatbots) awaiting signature alongside already-archived IL/OR/CA executive actions.

Evidence records
4
Cross-source patterns
1
Evidence classes
4 independent reporting
Outcomes
2 emerging1 effective1 cautionary
Source freshness
4 new this fortnight
Research completed
2026-09-26

Choose a role to see its takeaway beside every record in the ledger.

Patterns across the evidence

Synthesis · Lighthouse Advisory interpretation

1 pattern, each supported by at least two sources

  1. Digital States A-grade CIOs pair measured GenAI operations with governance-first scale stories

    In the 2026-09-23 A-state interview wave, Michigan publishes MDHHS document-AI and coding-assistant metrics with human eligibility retained, Nevada reports executive-branch Copilot plus call-center/cyber AI, and Georgia foregrounds Office of AI / Horizons Lab governance before tool scale—together showing the Digital States “infrastructure” narrative resolving into heterogeneous but concrete state operating choices.

    Operating questionWhen Digital States grades celebrate AI as infrastructure, which evidence should a state demand next—hours-saved IDP metrics, enterprise Copilot utilization, or lab graduation gates?

    Supporting evidenceMichigan Department of Technology, Management and Budget (DTMB) / Acting CIO Eric SwansonNevada Office of the CIO / Timothy GalluziGeorgia Technology Authority / CIO Shawnzia Thomas

Evidence ledger

Full record · every source keeps its link and limitations

4 records
  1. Independent reportingEffectiveNew this fortnight

    Michigan Digital States interview reports MDHHS document-AI hours saved and coding-assistant gains

    Why now: 2026-09-23 Digital States A-state interview (day before the archived Digital States 2026 overview/Utah Gemini piece) gives Michigan Acting CIO Eric Swanson’s measured GenAI results. MDHHS AI document processing for food-assistance renewals claims up to 97% automation on digital forms and up to 73% on handwritten forms, cutting an estimated 266 staff hours per week while eligibility decisions stay under program rules and staff review. DTMB’s AI coding assistant claims include updating .NET dependencies across 400+ files in two hours with zero errors, embedding WCAG 2.1 AA rules, rewriting classic ASP to .NET 10/Angular 21, and automating 165 of nearly 800 critical/high vulnerability remediations across 80 components. Separately, MSP Firearms Records Unit OCR claims ~241,800 records processed, ~42% auto-approval, ~101,500 manual reviews eliminated—public-safety adjacent; transfer only with stream limits.

    Michigan Department of Technology, Management and Budget (DTMB) / Acting CIO Eric SwansonMichigan, United States2026-09-23

    Why it matters, evidence and limitations
    Why it matters
    Concrete state HHS benefits-operations and enterprise IT productivity evidence from an A-grade CIO interview in the same Digital States wave as yesterday’s edition—directly transferable to other states’ SNAP/Medicaid document intake and SDLC copilots, with explicit human-in-the-loop eligibility language.
    Evidence and measured results
    GovTech Digital States Q&A quoting Swanson/DTMB program figures; operator-reported, not an independent statistical evaluation. MiCARA SNAP case-review tool described as supporting specialists. MSP firearms figures included in same answer but tagged as public-safety adjacent.
    Limitations and uncertainty
    Automation rates and hour-saved estimates are state-reported via journalism. MSP firearms OCR is public-safety/policing-adjacent—do not treat as core benefits evidence without stated limits. No baseline methodology or error-rate audit published here.
    Independent trade Q&A with A-grade state CIO in Digital States 2026 series Role takeaways & implementation
  2. Independent reportingEmergingNew this fortnight

    Nevada CIO cites executive-branch Copilot, AI call centers, and AI-enhanced cyber as top impacts

    Why now: 2026-09-23 Digital States A-state interview complements the prior day’s Digital States infrastructure narrative with Nevada-specific deployment claims. CIO Timothy Galluzi says the state launched Copilot across the executive branch and highlights AI tools supporting call centers plus AI enhancements to cybersecurity platforms and code review as the highest-impact AI work—arguing defenders must use AI because attackers do. He also ties IT priorities to 2025 special-session legislation enabling whole-of-state cybersecurity collaboration beyond the executive branch.

    Nevada Office of the CIO / Timothy GalluziNevada, United States2026-09-23

    Why it matters, evidence and limitations
    Why it matters
    State shared-services Copilot and contact-center AI claims from an A-grade CIO, useful for workforce-productivity and resident-service channel discussions, with explicit linkage to whole-of-state cyber operating models.
    Evidence and measured results
    GovTech Digital States Q&A; qualitative self-report without seat counts, deflection, or time-saved KPIs in the opened interview.
    Limitations and uncertainty
    No quantitative Copilot adoption or call-center outcome metrics in this interview; “amazing results” is qualitative. Distinct from Utah Gemini hours-saved piece archived 2026-09-24.
    Independent trade Q&A with A-grade state CIO in Digital States 2026 series Role takeaways & implementation
  3. Independent reportingEmergingNew this fortnight

    Georgia CIO emphasizes Office of AI and Horizons Innovation Lab as governance-first scale path

    Why now: 2026-09-23 Digital States A-state interview positions Georgia’s highest-impact AI work as institutional—Office of Artificial Intelligence plus Horizons Innovation Lab—rather than a single chatbot metric. CIO Shawnzia Thomas says Georgia led with governance, workforce education, and clear policies so agencies can explore AI securely, using the Lab as a “test kitchen” with academic and industry partners to validate value before scale. She links AI governance to cybersecurity principles (accountability, transparency, risk, trust) and describes a whole-of-state Security Operations Center and AI detection/response focus.

    Georgia Technology Authority / CIO Shawnzia ThomasGeorgia, United States2026-09-23

    Why it matters, evidence and limitations
    Why it matters
    Transferable state operating-model evidence for governments that need a governed path from pilots to enterprise adoption—complements Michigan’s measured IDP metrics and Nevada’s Copilot claims in the same interview wave.
    Evidence and measured results
    GovTech Digital States Q&A; qualitative CIO narrative. June 2026 Darwin partnership coverage exists elsewhere but is not the basis of this resource’s claims.
    Limitations and uncertainty
    No quantitative productivity or inventory KPIs in this interview; vendor partnerships not named here. Distinct from archived Georgia awards/NASCIO materials.
    Independent trade Q&A with A-grade state CIO in Digital States 2026 series Role takeaways & implementation
  4. Independent reportingCautionaryNew this fortnight

    Transparency Coalition Sept 25 tracker: 85 AI laws in 27 states; AB 1609 chatbot bill awaits signature

    Why now: same-day (2026-09-25) weekly tracker. TCAI reports Illinois, California, and Oregon governors took executive action this past week on frontier/AI oversight (details overlap EOs already archived 2026-09-23/24—treat as landscape wrap, not primary EO text). It points to TCAI’s mid-year report of 85 new AI-related laws passed in 27 states so far in 2026, lists six states still in session on AI bills, and enumerates California measures awaiting signature including AB 1609 on customer-service chatbots—relevant operating constraint for large private chatbots and a signal for resident-facing disclosure/human-access expectations.

    Transparency Coalition (AI Legislative Update)United States (multi-state; California focus on pending bills)2026-09-25

    Why it matters, evidence and limitations
    Why it matters
    Same-day multi-state legislative/oversight signal for state-government AI programs: volume of 2026 enactments and pending chatbot human-access rules that shape vendor and agency service design, with clear limits that EO specifics are already archived elsewhere.
    Evidence and measured results
    Advocacy weekly digest by Bruce Barcott; cites TCAI mid-year legislation report and bill lists. Not a government primary source; EO narratives are secondary to archived governor materials.
    Limitations and uncertainty
    Advocacy source; EO week wrap duplicates archived Illinois EO 2026-07, Oregon EO 26-26, and California N-9-26 expert/EO coverage—do not re-litigate those primary texts here. AB 1609 targets large private businesses, not state agencies, unless separately applied.
    Advocacy organization weekly state AI legislation digest Role takeaways & implementation

How to read this edition

Source findings, measured results and limitations come from the cited publications. Patterns, operating questions, role takeaways and implementation considerations are Lighthouse Advisory interpretation, stated as questions to validate locally rather than guaranteed outcomes. Vendor and operator claims are labeled as claims. Full research method.

Independent reporting
Independent reporting with attributable sources but without a formal evaluation design.