PLAYBOOK ATLASPlaybook Atlas

Public Sector

AI in public sector: explore 40 applications.

Help public servants review evidence and deliver services through accountable workflows.

Explore application intake, records research, caseworker assistance and infrastructure planning for public agencies.

Plan your first AI move See the 40 applications

Where the 40 applications sit

Explore the domains
  1. 23Government Administration
  2. 16Public Safety and Security
  3. 7Public Finance Management
  4. 2Infrastructure Management
  5. 1Economic Development

01Why AI now

Why public sector is moving on AI, in three figures.

Each figure links to its source.

  1. Government AI software: $70.6B by 2027

    Gartner forecast global government AI software spending would rise from $41.6 billion in 2024 to $70.6 billion in 2027.

    Source: Gartner

  2. Government improper payments: $186B in FY2025

    Fifteen U.S. federal agencies reported an estimated $186 billion in improper payments across 64 programs, and GAO says the total is incomplete.

    Source: U.S. Government Accountability Office

  3. 86% of OECD governments use AI internally

    The OECD says AI use for internal government processes rose from 70% of measured countries in 2023 to 86% in 2025.

    Source: OECD

How to evaluate a first project

Start with a specific service and jurisdiction. Identify the controlling sources, authorized reviewer and route for correction or appeal. Test missing records, accessibility and unequal error patterns, and evaluate completed service outcomes before expanding the role of an automated recommendation.

02Capability map

Where public sector puts AI to work, domain by domain.

Pick a domain to see its applications. Each one opens a full report.

Inside Government Administration

Applications in Government Administration

What the use cases ask AI to do

Share of the public sector use cases, by the kind of work the AI does.

  1. Reasoning

    Analyzes, predicts and decides: scoring, forecasting, classifying.

    60%

  2. Generation

    Creates: text, images, code and other content.

    33%

  3. Perception

    Sees, hears and reads: documents, images, audio and video.

    7%

Under 1% each: optimization and agentic.

03Workflow ideas

Ways AI could change the work, in public sector.

Atlas groups these applications by the changes they propose to a workflow. These preliminary classifications help you explore ideas; each application still needs a review of its sources and operating requirements.

Workflow categories in this catalog

Share of the 32 applications assigned a category.

How to evaluate an application

  1. Start with a bounded task

    Identify the inputs, the person responsible, and the decision or output the workflow must produce.

  2. Check the support for each method

    Look for source-backed details about the specific implementation, including its limits and human review.

  3. Measure the pilot

    Record the current baseline, agree on acceptance criteria, and test on representative cases before expanding.

Explore applications

04What gets built

The approaches public sector uses most, and when to skip them.

Knowing when not to use one matters as much as knowing when to.

  1. 01 · 9 applications

    Generative AI

    Models that learn data distributions to generate new text, images, code, and media.

    Use it when

    • When you need to generate natural language, code, or media that would be expensive or slow for humans to produce manually.
    • When tasks involve ambiguity, open-ended questions, or creative exploration rather than strict right-or-wrong answers.
    • When you can ground the model on reliable internal or external data sources to improve factuality.

    Skip it when

    • When the task requires strict determinism, formal guarantees, or regulatory-grade traceability (e.g., certain financial calculations, safety-critical control systems) without a robust verification layer.
    • When you lack access to high-quality, authoritative data or subject-matter expertise to validate outputs in a high-risk domain.
    • When a simple rules engine, search, or template-based system can solve the problem more reliably and cheaply.
    Generative AI, in depth

  2. 02 · 8 applications

    RAG-Standard

    LLM + retrieval layer that injects relevant documents into the prompt at query time.

    Use it when

    • You need an LLM to answer questions based on proprietary or private documents that are not part of its pretraining data.
    • Your knowledge changes frequently (e.g., policies, product docs, pricing) and you want updates to be reflected without re-training the model.
    • You have a large corpus of unstructured or semi-structured text (wikis, PDFs, tickets, emails, manuals) that users struggle to search effectively.

    Skip it when

    • Your task is primarily structured data querying or analytics (e.g., reporting over a relational database) where SQL or BI tools are more accurate and efficient.
    • You have very small, static knowledge (e.g., a short FAQ) that can fit directly into a prompt or be encoded as rules without a retrieval layer.
    • You require strict, deterministic behavior with no tolerance for probabilistic errors (e.g., certain financial calculations, safety-critical controls).
    RAG-Standard, in depth

  3. 03 · 4 applications

    Workflow Automation

    Use AI inside orchestrated workflows to automate multi-step business processes.

    Use it when

    • When you have repetitive, rule-heavy workflows that still require human judgment or document understanding.
    • When the process spans multiple systems (ERP, CRM, ticketing, email) and manual coordination is slow or error-prone.
    • When large volumes of semi-structured or unstructured data (emails, PDFs, forms) must be processed consistently.

    Skip it when

    • When the process is highly creative, exploratory, or non-repeatable, with no stable sequence of steps.
    • When decisions are extremely high-stakes (e.g., life-critical medical decisions, major legal judgments) and regulations require full human control.
    • When you lack reliable access to the necessary systems or data sources to complete the workflow end-to-end.
    Workflow Automation, in depth

05The applications

40 public sector applications, each with a full report.

The widest first, by the number of use cases each covers. Every report shows what changed, how it works, who has done it and what can go wrong.

  1. Program Integrity Fraud DetectionThis application area focuses on detecting, preventing, and managing fraud, waste, abuse, and corruption across government and quasi‑public programs, payments, and digital services. It encompasses benefits and claims fraud, procurement and supplier fraud, identity theft and account takeover, and broader financial crime affecting public funds. The core capability is to continuously monitor transactions, entities, and user behavior to flag anomalous patterns and prioritize high‑risk cases for investigation. It matters because traditional government fraud controls are largely manual, slow, and sample‑based, often catching issues only after funds are disbursed and hard to recover. By applying advanced analytics to large, heterogeneous datasets, organizations can shift from “pay and chase” to proactive prevention, reduce financial leakage, protect program integrity, and maintain public trust. At the same time, it helps governments respond to new threats such as AI‑enabled forgeries and at‑scale fraud campaigns by upgrading verification, oversight, and monitoring capabilities.Batch → RT22 use cases
  2. Public Sector Risk and Fraud Intelligence HubThis AI solution uses AI to predict crime hotspots, detect benefits and grant fraud, and surface emerging risks across public-sector programs. By combining geospatial analytics, bias-aware predictive policing, and advanced anomaly detection on financial and case data, it helps agencies target interventions, allocate resources, and reduce losses while improving community safety and trust.React → Pred13 use cases
  3. Smart City Service OrchestrationSmart City Service Orchestration is the coordinated use of data and automation to plan, deliver, and continually improve urban public services across domains such as transportation, energy, public safety, and citizen support. Instead of siloed, paper-heavy, and reactive departments, cities use integrated data and decision systems to route requests, prioritize interventions, and tailor services to different resident groups, languages, and accessibility needs. This turns fragmented digital touchpoints and back-office workflows into a single, responsive service layer for the city. AI is applied to fuse sensor, administrative, and citizen interaction data, predict demand, recommend actions to officials, and personalize information and service flows for individuals. It powers policy simulations, dynamic resource allocation, and automated handling of routine cases, while keeping humans in the loop for oversight and sensitive decisions. The result is faster responses, more inclusive access, better use of scarce budgets and staff, and a more transparent, trustworthy relationship between residents and local government.Silo → Int7 use cases
  4. Police Technology Governance MonitorPolice Technology Governance is the application area focused on systematically evaluating, regulating, and overseeing the use of surveillance, analytics, and digital tools in law enforcement. It combines legal, civil-rights, and policy analysis with data-driven insight into how policing technologies are acquired, deployed, and used in practice. The goal is to create clear, enforceable rules and oversight mechanisms that balance public safety objectives with privacy, equity, and constitutional protections. AI is applied to map and analyze patterns of technology adoption across agencies, surface risks (e.g., bias, over-surveillance, due-process issues), and generate evidence-based policy options. By mining procurement records, deployment data, usage logs, complaints, and case outcomes, these systems help policymakers, courts, and communities understand the real-world impacts of body-worn cameras, predictive tools, and other policing technologies. This supports the design of more precise regulations, accountability frameworks, and community oversight models. This application area matters because law enforcement agencies are rapidly adopting powerful technologies without consistent governance, exposing governments to legal liability, eroding public trust, and risking civil-rights violations. Structured governance supported by AI-driven analysis enables proactive risk management instead of reactive crisis response, and aligns technology deployments with democratic values and community expectations.Expert → AI6 use cases
  5. Public Investigations Case PrioritizationThis AI solution uses AI to support public‑sector investigations by detecting patterns of criminal activity, welfare fraud, and program abuse across diverse data sources. It prioritizes cases, flags high‑risk entities, and guides investigators with predictive insights, helping law enforcement and integrity bureaus focus resources where they are most likely to prevent crime, reduce fraud losses, and improve public safety.React → Pred5 use cases
  6. Predictive Patrol Allocation SystemPredictive policing is the use of data-driven models to forecast where and when crimes are likely to occur, and in some cases which individuals or groups are at higher risk of offending or victimization. By analyzing historical crime records, environmental factors, socioeconomic indicators, and real-time incident data, these systems generate risk scores, heatmaps, or priority lists that guide patrol routes, investigations, and preventive interventions. This application matters because police departments and public agencies operate under tight resource constraints while facing pressure to reduce crime, respond faster, and justify deployment decisions. Predictive policing promises more efficient use of officers and budgets, earlier intervention before crimes happen, and evidence-based planning for community programs. At the same time, it raises serious concerns about bias, transparency, legality, and public trust, driving parallel work on fairness assessment, bias detection, and governance frameworks for its responsible use.React → Pred5 use cases
All 40 applications, A to Z

Work with us

The catalog shows what is possible. A sprint decides what you should fund.

Reading every report still leaves the hardest question open: which move is worth funding first, here, given your constraints. That is the sprint: two weeks, fixed scope, and a candid answer either way.

Book a 30-minute callHow the sprint works

Two weeks Fixed scope A candid fit-or-no-fit answer

  1. 01

    Three ranked opportunities

    A decision-ready shortlist scored against value, feasibility, evidence, risk, and time to proof.

  2. 02

    One pilot-ready workflow

    A bounded operating design with actors, handoffs, data needs, controls, and success measures.

  3. 03

    A defensible 90-day plan

    Build-versus-buy guidance, ROI assumptions, risk treatment, sequencing, and executive decisions.

06Rules & standards

Rules and standards to check, with their scope and sources.

Selected US and EU references for these workflows. Applicability depends on jurisdiction, organization, data and intended use. The planning notes are Atlas editorial guidance; they are not official implementation-time estimates or an exhaustive list of obligations.

  1. OMB M-25-21 and M-25-22 · federal AI use and acquisition

    Agency guidance

    United States · covered federal agencies

    These memoranda address federal AI governance and acquisition. Scope and implementation depend on the agency and use; state and local governments have separate requirements.

    White House: federal agency AI use and procurement policies

    Source checked

    Project planning note

    Confirm the agency’s current policy, responsible officials and acquisition requirements for the proposed use.

  2. FedRAMP · agency use of cloud services

    Agency guidance

    United States · federal information systems using in-scope cloud offerings

    FedRAMP security evidence supports agency decisions, while the agency authorizes its own information system and accepts risk for the intended use. Confirm the offering and use are within the applicable boundary.

    FedRAMP: Using a certified cloud service

    Source checked

    Project planning note

    Check the exact cloud offering, service boundary and agency authorization path before selecting a deployment architecture.

Your next step

Find your first move in public sector.

Ask the workspace which move fits your company. Add your size, systems and constraints, and it works from these 40 applications to a plan your team can act on.

Plan your first AI move

Rather talk it through? Book a 30-minute call