Lighthouse AdvisorySLED AI Adoption Intelligence

Public Sector & Government · Issue 03 ·

State Government

Four unarchived records connect state AI testing, portfolio growth and oversight disagreement. Utah's historical tax pilot reports improved benchmark scores with test-reuse and production limits. New Zealand's August survey describes operational expansion without causal benefit evidence. Utah's regulator status and Medical Board letter preserve disputed consultation and incomplete benefit measurement, with the agency response inspected. Two cross-source interpretations support explicit acceptance gates. No fresh September 8 measured state-service improvement was established; independent current audit and clinical-outcome evidence remain gaps.

Evidence records
4
Cross-source patterns
2
Evidence classes
3 government evaluation1 standards or public-body guidance
Outcomes
2 mixed1 emerging1 cautionary
Source freshness
3 undated1 older, newly relevant
Research completed
2026-09-09

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

Synthesis · Lighthouse Advisory interpretation

Patterns across the evidence

2 patterns, each supported by at least two sources

  1. Expansion decisions need evidence beyond technical or adoption milestones

    Utah's tuned tax benchmark and NZ's operational-use counts answer different questions from improved service performance. Pair each continuation decision with representative workflow outcomes and human effort measures.

    Operating questionWhat observed quality and full-workflow effort measures would justify expanding this specific use?

    Supporting evidenceUtah Division of Technology Services and Utah State Tax CommissionNew Zealand Government Digital Delivery Agency

  2. Human review and institutional consultation need separate approval gates

    The Utah regulator describes case-level human review, while the Board challenges its own involvement before launch. An operational safeguard does not settle which institutions should advise or authorize the program.

    Operating questionWho must review the program before launch, and who can approve or stop each later increase in automation authority?

    Supporting evidenceUtah Office of Artificial Intelligence PolicyUtah Medical Licensing Board

Full record · every source keeps its link and limitations

Evidence ledger

4 records
  1. Government evaluationMixedUndated source

    Utah tax pilot improves benchmark scores, with generalization and production limits

    The state reports improved RAG answer scores after platform tuning; this is an operator evaluation, not proof of live service improvement.

    Utah Division of Technology Services and Utah State Tax CommissionUtah, United States2025 award submission; exact publication day unverified

    Why it matters, evidence and limitations
    Why it matters
    Direct state tax-assistance evidence, newly added as a historical procurement and validation case.
    Evidence and measured results
    July 2024–February 2025 project: four vendor platforms; 366 initial questions; expert rubric. Scores of 3 or 4 rose from 73% to 97%; top scores rose from 61% to 83%. Phase II reused lower-scoring questions, and one platform's testing stopped.
    Limitations and uncertainty
    No independent evaluation, held-out sample size or measured call-handling benefit established. Production was in progress when written; present status is unknown. Scores concern rubric categories, not universal accuracy. Chart text was inspected; screenshot yielded no inspectable image.
    State-authored NASCIO award submission Role takeaways & implementation
  2. Government evaluationEmergingUndated source

    New Zealand survey shows operational expansion while effectiveness remains self-reported

    Agency reporting indicates more operational AI use, but adoption counts do not establish causal service benefits.

    New Zealand Government Digital Delivery AgencyNew Zealand; transferable governance lessons for U.S. statesLast updated August 19, 2026

    Why it matters, evidence and limitations
    Why it matters
    A comparator for state shared-service portfolios, not a U.S. state census or a transferable legal framework.
    Evidence and measured results
    59 organisations reported 545 use cases, including 167 operational cases. The prior survey had 70 organisations and 272 cases. Benefits are agency-reported; changing respondents complicate trend interpretation.
    Limitations and uncertainty
    No controlled baseline, measured time savings, response-rate denominator or common outcome rubric supplied. August update is not a September measurement. Jurisdiction and survey composition limit transfer.
    Government survey report Role takeaways & implementation
  3. Government evaluationMixedUndated source

    Utah regulator keeps human review and acknowledges missing robust benefit evidence

    The public status page describes Phase 1 physician authorization and says robust benefit evidence is not yet available.

    Utah Office of Artificial Intelligence PolicyUtah, United StatesUndated status page inspected September 9, 2026 UTC

    Why it matters, evidence and limitations
    Why it matters
    State regulatory operating-model evidence; this is not a state-owned clinical deployment or an endorsement of the product.
    Evidence and measured results
    The page reports two company reports received and no serious incidents reported to the office. It describes adversarial vulnerabilities and changes to phase-transition criteria. These are regulator statements, not an independent safety finding.
    Limitations and uncertainty
    Undated relative timing cannot establish September operating status. Supporting May report relies on company physicians; independent review was initiated, not reported complete. That report's 'five months' heading conflicts with January–April wording; no duration or clinical accuracy percentage adopted.
    State regulator status and oversight page Role takeaways & implementation
  4. Standards or public-body guidanceCautionaryNewly relevant · Apr 2026

    Utah Medical Board objection exposes disagreement over pre-launch consultation

    The Board said it learned of the pilot after implementation and requested suspension pending discussion.

    Utah Medical Licensing BoardUtah, United StatesApril 20, 2026

    Why it matters, evidence and limitations
    Why it matters
    A state-agency consultation and decision-rights case, not proof of clinical harm.
    Evidence and measured results
    The April 20 letter is an oversight objection, not an outcomes study. Commerce's inspected April 21 response says medical experts reviewed the pilot and declines suspension, citing Phase 1 human review.
    Limitations and uncertainty
    A requested suspension is not an enacted suspension. The response is at https://commerce.utah.gov/wp-content/uploads/2026/04/Medical-Board-Doctronic-Response.pdf. Competing official positions are preserved; no adjudication of legal authority or current clinical safety is made.
    Primary oversight correspondence 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.

Government evaluation
A public body’s measured evaluation or documented pilot.
Standards or public-body guidance
Normative or advisory guidance from a standards body or public institution.