Nonprofit Technology Roadmap

What to Do When Your Board Wants an AI Roadmap

Your board isn't asking for a tool list. They're asking for confidence. Here's how to give them a roadmap that will actually work.

When a board asks for an AI roadmap, the instinct is to start listing tools: a chatbot for communications, predictive analytics for fundraising, a new dashboard for marketing.

"Your board isn't really asking for software recommendations. They're asking for confidence that the organization has a plan, isn't falling behind, and isn't creating avoidable risk."

That's a reasonable concern. The mistake is treating it like a shopping trip.

What Your Board Is Actually Asking

When a board asks for an AI roadmap, they're usually asking four practical questions:

  1. Where could AI actually help us?
  2. What risks do we need to manage?
  3. What should we do first, and what should wait?
  4. How do we know we're making responsible decisions?

If any of these feel fuzzy, you don't need a tool list yet. You need a decision framework.

A Better Approach: Build an AI Decision Plan

A useful AI roadmap isn't a list of tools. It's a sequence of outcomes that makes your organization more ready, more selective, and more responsible. It unfolds in four phases.

1AI ReadinessBefore any pilots

Before running pilots, make sure the organization is ready to use AI responsibly. Four things need to be in place:

  • Data Quality: Data needs to be usable, clearly defined, and accessible to the right people.
  • Governance Guardrails: Clear rules for privacy, sensitive information, and vendor review must be set before experimentation.
  • Clear Ownership: Decide who approves AI use cases, who reviews risk, and who is accountable for outcomes.
  • Staff Training: People need to understand where AI helps, where it falls short, and where human review is still required.
2Use Case DiscoveryFind the right opportunities

This is where most AI roadmap conversations should start, but rarely do. Look for repetitive work, decisions made with incomplete information, tasks where staff are writing the same thing repeatedly, and high-volume questions that are safe to automate.

Target outcomes:

  • Identify and score 10–15 candidate AI use cases across departments
  • A clear owner who is accountable for the outcome
3Scoped PilotsSmall, bounded, measurable

Your first pilots should be narrow. Not a chatbot that answers everything, more like one that answers two questions well, with clear guardrails. A good pilot has:

  • Narrow scope with defined boundaries
  • A clear owner who is accountable for the outcome
  • Explicit data boundaries and privacy rules
  • Measurable success signals you can evaluate
  • An exit plan if the pilot isn't working
4Scale What WorksReal ownership, real budget

If a pilot works, scale it, with real ownership, documentation, and budget attached. If it doesn't, capture the learning and stop.

A Roadmap is a Planning Process

Building an AI roadmap is real work. It requires time for use case review, policy decisions, data assessment, staff input, and pilot design.

If the board wants a credible roadmap, the organization has to resource that planning work. Without the time, ownership, and capacity to do it well, the roadmap may look finished long before the organization is actually ready.

Next Step

Start With a Plan, Not a Tool List

We help nonprofit leadership teams respond to AI pressure with a practical plan, not a rushed tool list.

Book a Strategy Conversation