Loading...
Loading...
New Hampshire state government is genuinely small by most measures — 1.4 million residents, a General Court legislature of 424 members (the third-largest legislative body in the English-speaking world by size relative to constituents), and a state budget that operates without a personal income tax or sales tax. That fiscal structure creates a government AI environment shaped by resource constraints that are philosophically different from those of larger states: New Hampshire isn't choosing between competing AI priorities; it's often choosing between AI investment and baseline service capacity. The Department of Information Technology (DoIT) is the enterprise IT authority, operating shared infrastructure across executive agencies on an Azure Government foundation, but with a team of fewer than 150 IT staff statewide — meaning AI projects that require sustained DoIT engineering support are competing with basic infrastructure maintenance for bandwidth. What New Hampshire does have that its size doesn't predict is a talent spillover from the Nashua-to-Manchester corridor: BAE Systems (the state's largest private employer, with 6,000+ employees at its Nashua facilities), Liberty Mutual's 2,700-person Manchester operations center, and Fidelity Investments' 4,000-employee Merrimack campus create a private-sector AI talent ecosystem that is dense for a state this size. Liberty Mutual's actuarial analytics team — which runs sophisticated ML models for risk pricing and claims prediction — produces professionals who have taken positions in the New Hampshire Insurance Department and the Department of Revenue Administration, bringing private-sector AI fluency into state government at levels that don't show up in staffing charts. The Department of Revenue Administration (DRA), which administers the Business Profits Tax, Business Enterprise Tax, and Meals & Rentals Tax in a state with no income tax, has been the lead state agency on AI adoption precisely because its revenue-maximization mandate creates clear ROI metrics for tax-gap analytics.
Updated June 2026
New Hampshire DoIT is organized into functional bureaus — Bureau of Client Services, Bureau of Infrastructure, Bureau of Information Security — rather than agency-aligned IT teams, which means state agencies don't have embedded IT staff and must work through DoIT project managers for any technology implementation above basic software configuration. For AI deployments, this creates a queue-based dynamic: DoIT's project-management capacity is the bottleneck, not procurement authority or budget. Agencies that have successfully deployed AI in New Hampshire have consistently used a co-delivery model: a vendor handles the AI development and integration, DoIT handles security review and infrastructure provisioning, and the agency handles change management and training — with formal roles documented in a project charter before work begins. DoIT's Information Security Policy, revised in 2024, incorporates NIST AI RMF criteria for AI systems classified as high-risk under state guidelines. New Hampshire's small size means DoIT's Bureau of Information Security Director has reviewed every significant AI deployment personally — an unusual level of CISO engagement that creates consistency but also creates a single-point-of-review bottleneck. The New Hampshire Information Technology Council, a legislative advisory body, published recommendations in 2024 calling for a dedicated AI governance coordinator within DoIT — a position funded in the FY2025-2026 budget at $145,000, representing the state's first explicit AI governance investment. The practical guidance for AI vendors in this market: shorter, staged projects with clear DoIT infrastructure-provisioning milestones outperform large-scope implementations that require sustained DoIT engineering involvement across 12+ months.
New Hampshire's Department of Revenue Administration operates in a revenue environment unlike any other state: it collects no personal income tax and no sales tax, making its primary revenue sources the Business Profits Tax (BPT), Business Enterprise Tax (BET), Meals and Rentals Tax (MRT), Real Estate Transfer Tax (RETT), and a small set of excise taxes. This structure makes DRA's compliance analytics function disproportionately important — a 1% improvement in BPT audit selection accuracy has larger dollar impact than in states with broad-based income tax because the business-tax base is more concentrated. DRA deployed ML-assisted audit selection for BPT returns in 2022, using a model trained on five years of audit history, industry-specific revenue and expense ratio benchmarks, and entity-relationship patterns (common ownership, shared addresses, related-party transactions) that flag structuring arrangements. The model replaced a largely manual selection process that had been unchanged since the 1990s. DRA reported a 31% improvement in audit yield per FTE in its FY2023 annual report — audits selected by the ML model produced higher average assessments than manually selected audits at the same staffing level. The Meals and Rentals Tax presents a separate AI application: with Strafford County and Rockingham County generating significant MRT revenue from Portsmouth's restaurant scene, Hampton Beach's seasonal hospitality, and the Manchester-to-Nashua restaurant corridor, ML-assisted MRT compliance monitoring uses point-of-sale data (from state-registered terminal providers) and benchmark comparisons to flag reporting anomalies. DRA's partnership with the New Hampshire Lodging and Restaurant Association has helped calibrate seasonal benchmarks — the Hampton Beach summer peak creates MRT filing patterns that flat-benchmark models consistently misinterpret without seasonal adjustment.
New Hampshire's government AI capabilities are disproportionate to its size partly because of talent that migrates from its largest private employers. Liberty Mutual's Manchester operations center — one of its two largest domestic hubs — runs sophisticated actuarial ML models for commercial lines pricing, claims prediction, and fraud detection. The New Hampshire Insurance Department, which regulates Liberty Mutual, Anthem, Harvard Pilgrim, and dozens of smaller carriers writing $9+ billion in annual premiums in the state, has benefited from staff who came from Liberty Mutual or other carrier analytics teams and understand ML model documentation, actuarial assumption transparency, and claims-pattern analysis at a level that most state insurance departments lack. The Insurance Department's AI review framework — used to evaluate carrier AI-based underwriting and pricing models submitted for rate approval — is modeled on the NAIC's 2023 AI/ML regulatory framework but with NH-specific adjustments reflecting the state's concentration in commercial-lines property insurance and the particular pricing volatility of the New Hampshire coastal market (Seacoast region, including Hampton and Portsmouth). BAE Systems' Nashua facilities develop electronic warfare systems and advanced defense sensors, creating a population of engineers with signal-processing, anomaly-detection, and sensor-fusion backgrounds who represent a talent pool for state AI projects in areas like environmental monitoring, transportation analytics, and emergency communications. The New Hampshire Bureau of Emergency Communications, which operates the state's Enhanced 911 system, has informally consulted with BAE engineers on next-generation dispatch analytics. DEKA Research & Development in Manchester — Dean Kamen's R&D firm — has also been a resource for state government technology problem-solving, particularly for the Department of Health and Human Services on remote-patient-monitoring programs for the state's rural population.
Strategic planning for AI adoption, readiness assessment, and roadmap development
Workflow automation using AI, including Make.com-style automation and RPA
Predictive models, data analysis, and ML pipeline development
Text analysis, document automation, sentiment analysis, and language processing
Without personal income tax, New Hampshire's revenue base is narrower and more concentrated in business taxes, rooms/meals taxes, and excise taxes — meaning DRA's compliance analytics have outsized revenue impact compared to most states. A 31% improvement in BPT audit yield per FTE, as DRA reported in FY2023, translates directly to tens of millions in additional revenue in a state with a $1.8 billion general fund. This ROI clarity has made DRA the strongest internal advocate for AI investment and the clearest internal proof-of-concept that other NH agencies reference when making AI budget cases.
DoIT's total IT staff is fewer than 150 people serving all executive-branch agencies — a staffing ratio that makes sustained engineering involvement in vendor-led AI projects genuinely scarce. Successful NH government AI projects use a co-delivery model: the vendor owns development and integration, DoIT provides security review (Bureau of Information Security) and Azure Government provisioning (Bureau of Infrastructure), with clearly scoped deliverables that don't assume DoIT engineering backfill. Projects scoped to complete DoIT security review within 6-8 weeks (using the NIST AI RMF assessment checklist DoIT published in 2024) move to production in 5-8 months total. Projects that assume ongoing DoIT engineering support routinely run 12-18 months.
The NH Insurance Department's AI review framework — based on the NAIC 2023 AI/ML Regulatory Guidance — requires carriers submitting AI-based rate or underwriting models to provide model documentation including training data description, feature importance analysis, disparate-impact testing results, and performance metrics on New Hampshire-specific test populations. The Department's actuarial staff, several of whom have Liberty Mutual or Anthem analytics backgrounds, conduct independent validation rather than accepting vendor self-certification. Carriers using AI for commercial-lines property pricing in the Seacoast region must document hurricane and coastal-storm assumptions specifically, given regulatory sensitivity about coastal availability issues that emerged after the 2012 Sandy storm response.
New Hampshire government AI projects are among the lowest-cost in the Northeast, primarily because the state's lean government structure means smaller integration footprints and lower-volume processing requirements. A mid-scale ML compliance-analytics project (similar to DRA's BPT audit selection model) runs approximately $150,000-$350,000 inclusive of DoIT security review, Azure Government integration, and 12 months of model monitoring. The constraint is not cost but DoIT project-queue capacity — even well-funded projects wait 4-6 months for DoIT infrastructure provisioning slots in the current environment.
New Hampshire DHHS, which manages Medicaid, behavioral health, and family services for a population spread across dense southern tier cities and very sparse northern tier towns, has focused AI investment on telehealth coordination and case-management triage rather than self-service automation. A 2023 DHHS pilot with Dartmouth Health (formerly Dartmouth-Hitchcock Medical Center) on AI-assisted behavioral health risk stratification for Medicaid enrollees produced a triage model that identifies high-utilization-risk individuals for care-management outreach — an application designed for caseworker decision support rather than automated determination, reflecting the political sensitivity around automated benefits decisions in a state with a libertarian political culture.