Playbook

AI Readiness Playbook

A practical guide for mission-driven organizations evaluating AI. No hype, no jargon — just a clear framework for understanding what AI can do for your organization, whether you're ready, and how to get started.

AI won't replace your mission. But organizations that adopt it thoughtfully will serve their constituents faster, more equitably, and at lower cost.

What's Inside

AI in Government & Nonprofits: What's Real

Let's cut through the noise. AI is not going to replace government workers. It's not going to solve homelessness or fix your permitting backlog overnight. But it is genuinely useful for specific, well-defined tasks — and organizations that ignore it will fall behind.

Here's what AI actually does well today for mission-driven organizations:

Document Processing

Production-ready

Extracting data from forms, permits, applications, and correspondence. AI can read handwritten text, classify documents, and route them to the right department — turning days of manual processing into minutes.

Constituent Services

Production-ready

Chatbots and virtual assistants that handle routine inquiries: office hours, application status, how-to questions. Not replacing human agents — handling the 60% of questions that have straightforward answers so staff can focus on complex cases.

Data Analysis & Reporting

Production-ready

Analyzing large datasets to identify trends, anomalies, and patterns. Budget forecasting, grant impact measurement, service demand prediction. AI finds what humans would take weeks to uncover.

Content Generation & Translation

Production-ready

Drafting public communications, translating content into multiple languages, generating meeting summaries. Always with human review — AI as a first draft tool, not a publisher.

Predictive Maintenance

Emerging

For organizations managing physical infrastructure: predicting equipment failures, optimizing maintenance schedules, reducing downtime. Requires historical maintenance data.

Policy Analysis

Emerging

Summarizing legislation, comparing policy options, analyzing public comments. Useful for research and preparation — not for making policy decisions.

The AI Readiness Assessment

Before investing in AI, assess your organization across five dimensions. Be honest — the goal isn't to score well, it's to understand where you need to invest before AI can deliver value.

1

Data Quality

AI is only as good as the data it learns from. If your data is inconsistent, incomplete, or siloed, AI will amplify those problems, not solve them.

Key Questions

  • Is your data digitized and structured?
  • Do you have consistent data standards across departments?
  • Is data regularly cleaned, validated, and updated?
  • Can you access and combine data across systems?
  • Do you know what data you have and where it lives?

Maturity Levels

Low Data is mostly in PDFs, spreadsheets, and email. No standards.
Medium Key data is in databases with some standards. Some silos remain.
High Clean, structured data with consistent standards and APIs. Data catalog exists.
2

Technical Infrastructure

AI tools need modern infrastructure to run on — APIs, cloud services, and integration capabilities.

Key Questions

  • Are your core systems accessible via APIs?
  • Do you use cloud services (or have a path to them)?
  • Can you provision new tools and services without months of procurement?
  • Is your network and security infrastructure compatible with AI services?
  • Do you have development and staging environments for testing?

Maturity Levels

Low Legacy systems with no APIs. On-premise only. Long procurement cycles.
Medium Mix of legacy and modern. Some cloud services. APIs for key systems.
High Cloud-native or cloud-ready. Modern APIs. CI/CD pipelines. Flexible procurement.
3

Team Capabilities

You don't need a team of data scientists. You need people who can evaluate AI tools, manage implementations, and interpret results.

Key Questions

  • Does anyone on your team understand AI fundamentals?
  • Do you have staff who can evaluate vendor AI claims critically?
  • Can your IT team integrate third-party APIs?
  • Do you have project management capacity for AI pilots?
  • Is there appetite for experimentation and learning?

Maturity Levels

Low No AI knowledge. Limited technical staff. No experimentation culture.
Medium Some team members are AI-curious. Capable IT staff. Open to pilots.
High AI-literate leadership. Strong technical team. History of successful innovation.
4

Organizational Culture

AI adoption fails more often for organizational reasons than technical ones. Change management matters more than the technology.

Key Questions

  • Is leadership actively supportive of AI exploration?
  • Are staff open to changing how they work?
  • Does your organization tolerate experimentation and failure?
  • Is there trust between departments to collaborate on cross-functional projects?
  • Do you have a history of successful technology adoption?

Maturity Levels

Low Risk-averse culture. Siloed departments. History of failed tech projects.
Medium Leadership interested but cautious. Some cross-functional collaboration.
High Innovation-oriented culture. Strong cross-functional collaboration. Proven change management.
5

Governance & Ethics

Government and nonprofit organizations have unique obligations around transparency, equity, and accountability that AI makes more complex.

Key Questions

  • Do you have a data governance framework?
  • Have you considered AI ethics policies?
  • Do you understand your obligations around algorithmic transparency?
  • Can you explain AI-assisted decisions to constituents?
  • Do you have processes for auditing AI outputs for bias?

Maturity Levels

Low No data governance. No AI policies. Haven't considered ethics implications.
Medium Basic data governance. Aware of ethical concerns. No formal AI policies yet.
High Strong data governance. AI ethics framework. Bias auditing processes. Transparency commitments.

Finding Your First AI Win

Your first AI project sets the tone for everything that follows. Choose wisely. A successful pilot builds organizational confidence and executive support. A failed pilot — even if the failure was due to poor project selection, not poor technology — can set your AI efforts back years.

Evaluate potential pilot projects against these six criteria:

High Volume

The task is performed hundreds or thousands of times. Even small efficiency gains multiply into significant impact.

Structured Process

The task follows clear rules or patterns. AI excels at consistent, repeatable work — not judgment calls in ambiguous situations.

Measurable Outcome

You can define success in numbers: time saved, accuracy improved, cost reduced. If you can't measure it, you can't prove the pilot worked.

Low Risk

If the AI gets it wrong, the consequence is an inconvenience, not a crisis. Save high-stakes decisions for after you've built confidence.

Available Data

You have the historical data needed to train or evaluate the AI tool. No data = no AI project.

Willing Team

The people doing this work today are open to the pilot. Forcing AI on a resistant team is a recipe for failure.

A project that scores well on all six criteria is a strong pilot candidate. If it only hits 3–4, think carefully about whether now is the right time.

Build vs. Buy vs. Partner

There are three paths to AI adoption. The right choice depends on your technical capacity, budget, timeline, and the complexity of your needs.

Off-the-Shelf Tools

$0–$500/month

Best when: Your need is common and well-served by existing products

Examples: ChatGPT/Claude for drafting, Google Translate for basic translation, Grammarly for content editing, pre-built chatbot platforms

Advantages

  • + Fast to deploy
  • + Low upfront cost
  • + Vendor handles maintenance
  • + Proven technology

Considerations

  • − Limited customization
  • − Data privacy concerns
  • − Vendor lock-in risk
  • − May not meet government compliance requirements

Custom Build

$50K–$500K+

Best when: Your needs are unique and you have strong technical capacity

Examples: Custom document processing pipelines, specialized prediction models, domain-specific classification systems

Advantages

  • + Tailored to your needs
  • + Full control over data
  • + No vendor dependency
  • + Can meet specific compliance requirements

Considerations

  • − High upfront cost
  • − Requires technical expertise
  • − Ongoing maintenance burden
  • − Longer time to deploy

Implementation Partner

$25K–$200K

Best when: You need customization but don't have in-house AI expertise

Examples: AI systems audit and strategy, custom chatbot implementation, data pipeline development, AI governance framework

Advantages

  • + Expert guidance
  • + Faster than building alone
  • + Knowledge transfer to your team
  • + Right-sized solutions

Considerations

  • − External dependency during implementation
  • − Need to evaluate partner carefully
  • − Requires internal project management

Governance & Ethics

Government and nonprofit organizations serve the public trust. That means AI adoption comes with obligations that private companies don't have. Getting governance right isn't a barrier to AI adoption — it's what makes adoption sustainable.

Bias and Equity

AI systems can perpetuate and amplify existing biases in your data. If historical permitting data shows disparities by neighborhood (which often correlates with race and income), an AI trained on that data will replicate those disparities. Before deploying any AI system that affects constituents, audit it for bias across protected categories.

Transparency

Constituents have a right to know when AI is involved in decisions that affect them. At minimum, disclose when AI is used in decision-making, explain what factors the AI considers, and provide a human review option for consequential decisions. "The algorithm decided" is never an acceptable answer.

Data Privacy

Understand where your data goes when you use AI tools. Many commercial AI services use customer data for training. For government data — especially PII — this may violate your privacy obligations. Negotiate data processing agreements. Consider on-premise or government-cloud AI options for sensitive data.

Constituent Trust

Public trust is hard to build and easy to lose. A single AI mistake that goes viral — a biased chatbot response, a wrongful benefit denial — can undermine years of digital trust-building. Start with low-risk applications, build in human oversight, and be transparent about what AI is and isn't doing.

Policy Framework

Adopt an AI policy before deploying AI. Key elements: approved use cases, prohibited uses, data governance requirements, bias testing protocols, human oversight requirements, transparency commitments, and incident response procedures. Several cities (San José, Seattle, Boston) have published AI policies that can serve as templates.

Implementation Roadmap: 90-Day Plan

Twelve weeks to go from exploration to evidence. This roadmap moves fast enough to maintain momentum but slow enough to make thoughtful decisions.

Weeks 1–4

Assess

  • Complete the AI Readiness Assessment (all 5 dimensions)
  • Inventory current data assets and quality
  • Identify 3–5 potential pilot use cases
  • Research what peer organizations are doing with AI
  • Brief leadership on findings and recommendations
  • Select one pilot project using the criteria framework
Weeks 5–8

Pilot

  • Define success metrics for the pilot
  • Select tools or partner for implementation
  • Set up data pipeline and testing environment
  • Run the pilot with a small group
  • Collect feedback from users and stakeholders
  • Document results, challenges, and lessons learned
Weeks 9–10

Evaluate

  • Measure pilot results against success criteria
  • Calculate ROI and project future savings at scale
  • Gather qualitative feedback from staff and constituents
  • Identify what worked, what didn't, and what surprised you
  • Assess readiness for scaling
  • Present findings to leadership with recommendation
Weeks 11–12

Scale or Pivot

  • If successful: plan phased rollout to broader organization
  • If mixed results: identify adjustments and run a second iteration
  • If unsuccessful: document learnings and select next pilot candidate
  • Develop AI governance policy based on pilot experience
  • Create training materials for broader team
  • Build the business case for the next AI investment

What It Costs

Honest budget ranges for different types of AI projects. These reflect what we've seen across government and nonprofit organizations, including implementation, training, and initial support.

AI Strategy & Assessment

$10K–$30K 2–4 weeks

Understanding where you are and where to start. Includes readiness assessment, use case identification, and roadmap.

Chatbot / Virtual Assistant

$25K–$100K 6–12 weeks

From simple FAQ bot to a multi-channel assistant integrated with your systems. Cost depends on complexity and integrations.

Document Processing Automation

$30K–$150K 8–16 weeks

Automating intake, classification, and data extraction from forms and documents. Requires training data and integration work.

Data Analysis & Dashboards

$20K–$80K 4–10 weeks

AI-powered analytics and reporting. Includes data cleaning, model development, and visualization.

Custom AI Application

$75K–$500K+ 3–12 months

Purpose-built AI solution for your specific needs. Significant investment but can deliver transformative results.

Budget Planning Tips

  • • Budget 15–20% of implementation cost for annual maintenance and improvements
  • • Include training costs — technology without adoption is wasted money
  • • Start with assessment before committing to implementation budget
  • • Factor in staff time — your team will spend significant time on any AI project

Ready to assess your AI readiness?

Our AI Systems Audit gives you a clear picture of where you stand, what's possible, and a prioritized roadmap to get there. No vendor pitches — just honest guidance.

Learn About Our AI Systems Audit