Executive summary

AI can’t fix broken processes—it amplifies them. Local governments that standardize workflows, classify data, and embed GRC before piloting agents or copilots will see faster ROI and fewer surprises. CSF 2.0, NIST’s AI RMF (with a Generative AI Profile), ISO/IEC 42001, and CISA’s Cybersecurity Performance Goals give you a clear backbone. Microsoft Purview provides pragmatic controls you can turn on now. CISA+4NIST+4NIST Publications+4

Why now

• CSF 2.0 adds a “Govern” function—a cue to treat cybersecurity and risk as enterprise‑wide governance, not just IT plumbing. NIST
• Agentic AI is surging—but many projects will be cut. Gartner (via Reuters) expects >40% of agentic AI projects to be scrapped by 2027 over cost, unclear value, and immaturity. Translation: only governed use‑cases will endure. Reuters
• MODPA takes effect Oct 1, 2025. Even where counties aren’t directly covered, your vendors and data‑sharing partners likely are—so harmonizing classification, retention, and sharing rules reduces risk. Maryland General Assembly+1

DOPI’s thesis

The fastest path to resilient, AI‑ready government is a tight braid of:

1. Process excellence (documented value streams + SOPs),
2. Embedded GRC (controls mapped to standards), and
3. Data governance (classification, retention, DLP)—
   implemented on platforms you already own. CSF 2.0 + NIST AI RMF/GenAI Profile + ISO/IEC 42001 are the north stars; CISA’s CPGs and Zero Trust guidance give you milestones. The White House+4NIST+4NIST Publications+4

What “good” looks like (framework‑to‑control map)

• Cyber risk governance: Anchor program strategy and metrics to NIST CSF 2.0. Use the Govern function to connect risk decisions to business outcomes. NIST
• AI governance: Stand up policies and evaluation criteria per NIST AI RMF and the Generative AI Profile; formalize management system practices with ISO/IEC 42001. NIST+2NIST Publications+2
• Zero Trust: Align with OMB M‑22‑09 and CISA’s Zero Trust Maturity Model guidance. The White House+1
• Business continuity: Map continuity objectives to ISO 22301. ISO
• Data classification & retention: Use NIST SP 800‑60 for information types and NARA GRS 4.2 for retention/access records. NIST Publications+2NIST Publications+2
• Incident readiness: Update playbooks to NIST SP 800‑61 Rev. 3 (2025). NIST Publications+1

Agentic AI for GRC: where it actually helps

Start where value is clear and risk is low: policy routing, evidence gathering, alert triage, and audit preparation. Keep humans in the loop; measure outcomes. Expect market churn: more than 40% of agentic initiatives may be canceled by 2027, so insist on KPIs and exit criteria. Reuters

A 90‑day playbook (built around tools you already have)

Days 0–30 — Baseline & blueprint

• Map 2–3 priority services (e.g., permits, inspections) and write the “happy path + exceptions” as SOPs.
• Draft a classification schema (NIST 800‑60) and a retention crosswalk (NARA GRS 4.2). NIST Publications+1
• In Microsoft Purview, define sensitivity labels and auto‑labeling rules; plan DLP pilot scope. Microsoft Learn+2Microsoft Learn+2

Days 31–60 — Controls & pilots

• Publish SOPs with embedded Zero Trust checkpoints tied to CSF 2.0 categories. The White House+1
• Enable Purview DLP in priority locations (Exchange, SharePoint/OneDrive, endpoints). Start DLP for Microsoft 365 Copilot location for “high‑risk” units. Microsoft Learn+1
• Approve 1–2 AI/GRC pilot use‑cases with AI RMF measures (evaluation criteria, human oversight, incident procedures). NIST

Days 61–90 — Resilience & scale

• Update incident response per NIST SP 800‑61 Rev. 3; run a tabletop. NIST Publications
• Align continuity artifacts to ISO 22301 (RTO/RPO, comms trees, supplier dependencies). ISO
• Stand up a KPI dashboard (see below) and schedule quarterly CSF 2.0 governance reviews. NIST

Microsoft Purview: quick‑wins checklist

• Sensitivity labels created, ordered, and published (e.g., Public / Personal/ General et.). Microsoft Learn
• Auto‑labeling for high‑value SITs (SSNs, payment data, health identifiers). Microsoft Learn+1
• DLP policies for data‑at‑rest, in‑use, and in‑motion across Exchange, SharePoint/OneDrive, Teams, endpoints. Microsoft Learn
• DLP for Microsoft 365 Copilot policy location configured (block processing of restricted items; enable alerting). Microsoft Learn

KPIs that matter (manage what you can prove)

• Process coverage: % priority services with SOPs + mapped controls (CSF 2.0 categories). NIST
• Data policy drift: DLP violations per 1,000 users; mean time to remediate (Purview). Microsoft Learn
• Incident performance: MTTD/MTTR vs. SP 800‑61 r3 playbooks; tabletop cadence. NIST Publications
• Continuity readiness: RTO/RPO attainment and % of critical processes tested (ISO 22301). ISO
• AI safety & value: # tasks automated, human‑in‑the‑loop overrides, eval results per AI RMF. NIST

---

FAQs:

Q1: Are counties and cities directly subject to MODPA? MODPA primarily applies to businesses, but because many public entities share data with covered partners or vendors, aligning your classification, retention, and sharing practices reduces friction and risk across the ecosystem. Maryland General Assembly+1

Q2: Isn’t Zero Trust just a federal thing? No. OMB M‑22‑09 mandates it for federal agencies, but the principles and CISA’s maturity guidance are broadly useful for any government enterprise modernizing identity, data, and network controls. The White House+1

Q3: What’s the safest starting point for “agentic AI” in government? Low‑risk GRC workflows (control evidence prep, policy routing, alert triage) with AI RMF guardrails, while the market matures—especially given Gartner’s cancellation forecast. NIST+1

---

How this ties to DOPI’s mission

DOPI helps counties and cities braid process excellence, embedded GRC, and data governance—leveraging Microsoft’s stack—to achieve safer, smarter, more resilient operations. Our approach is standards‑anchored (NIST/CISA/ISO) and outcome‑driven (90‑day playbooks, measurable KPIs).

Schedule a 45‑minute readiness review. We’ll map one service, one risk, and three data controls you can turn on this quarter.

---

Sources & further reading:

• National Institute of Standards and Technology. (2024, February 26). The NIST Cybersecurity Framework (CSF) 2.0 (NIST Cybersecurity White Paper NIST.CSWP.29). https://nvlpubs.nist.gov/nistpubs/CSWP/NIST.CSWP.29.pdf NIST Publications
• National Institute of Standards and Technology. (2023, January). Artificial Intelligence Risk Management Framework (AI RMF 1.0) (NIST AI 100-1). https://nvlpubs.nist.gov/nistpubs/ai/nist.ai.100-1.pdf NIST Publications
• National Institute of Standards and Technology. (2024, July 25). Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile (NIST AI 600-1). https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.600-1.pdf NIST Publications
• International Organization for Standardization. (2019). ISO 22301:2019—Security and resilience—Business continuity management systems—Requirements. https://www.iso.org/standard/75106.html ISO
• International Organization for Standardization. (2023, December). ISO/IEC 42001:2023—Information technology—Artificial intelligence—Management system. https://www.iso.org/standard/42001 ISO
• Cybersecurity and Infrastructure Security Agency. (n.d.). Cybersecurity Performance Goals (CPGs): Frequently asked questions. (Aligned to NIST CSF functions.) Retrieved September 11, 2025, from https://www.cisa.gov/cross-sector-cybersecurity-performance-goals/frequently-asked-questions CISA
• Office of Management and Budget. (2022, January 26). Moving the U.S. Government toward zero trust cybersecurity principles (Memorandum M‑22‑09). https://www.whitehouse.gov/wp-content/uploads/2022/01/M-22-09.pdf The White House
• Stine, K., Kissel, R., Barker, W., Fahlsing, J., & Gulick, J. (2008). Guide for mapping types of information and information systems to security categories (NIST SP 800‑60, Vol. 1, Rev. 1). National Institute of Standards and Technology. https://csrc.nist.gov/pubs/sp/800/60/v1/r1/final NIST Computer Security Resource Center
• National Archives and Records Administration. (2023, June 2). General Records Schedule 4.2: Information access and protection records (Transmittal 34). https://www.archives.gov/files/records-mgmt/grs/grs04-2.pdf National Archives
• Nelson, A., Rekhi, S., Souppaya, M., & Scarfone, K. (2025, April). Incident response recommendations and considerations for cybersecurity risk management: A CSF 2.0 community profile (NIST SP 800‑61 Rev. 3). National Institute of Standards and Technology. https://doi.org/10.6028/NIST.SP.800-61r3 NIST Computer Security Resource Center
• Reuters. (2025, June 25). Over 40% of agentic AI projects will be scrapped by 2027, Gartner says. https://www.reuters.com/business/over-40-agentic-ai-projects-will-be-scrapped-by-2027-gartner-says-2025-06-25/ Reuters
• Maryland General Assembly. (2024, May 9). Maryland Online Data Privacy Act of 2024 (HB 567), Chapter 454 (effective October 1, 2025). https://mgaleg.maryland.gov/2024RS/Chapters_noln/CH_454_hb0567e.pdf Maryland General Assembly
• Microsoft. (2025, July 17). Learn about sensitivity labels (Microsoft Purview). https://learn.microsoft.com/en-us/purview/sensitivity-labels Microsoft Learn
• Microsoft. (2025, July 7). Automatically apply a sensitivity label to Microsoft 365 data (auto‑labeling). https://learn.microsoft.com/en-us/purview/apply-sensitivity-label-automatically Microsoft Learn
• Microsoft. (2025, June 5). Learn about data loss prevention (Microsoft Purview DLP). https://learn.microsoft.com/en-us/purview/dlp-learn-about-dlp Microsoft Learn
• Microsoft. (2025). Learn about the Microsoft 365 Copilot policy location (DLP). https://learn.microsoft.com/en-us/purview/dlp-microsoft365-copilot-location-learn-about Microsoft Learn