• A clear picture of who you’re actually solving for, based on real people not assumptions
• A shortlist of the problems in your organization where AI is most likely to help
• A way to assess which ideas are worth pursuing and which aren’t ready yet
• A concrete 7-day test plan you can take to your leadership team

• Work through it with your program team over 2 to 3 half-day workshops (or one focused week).
• Follow the sections in order. Each builds on the last.
• Use the Notion template alongside it as your team’s live working document.
• When you finish, test your top use case with the 7-Day Mini-Pilot before committing further resources.

You can engage with this playbook directly on NotebookLM, an AI-powered notebook that lets you ask questions in plain language and get answers drawn only from the playbook content, nothing outside it.

Use the prompts below to walk through the method with your own organization’s context. Follow them in order for the best results.

Prompt 1: “We serve [describe your population] through [describe your program]. Walk me through how to build a beneficiary persona and journey map using the template in Section 2. What should I include beyond basic demographics?”

Prompt 2: “Here are three pain points my team identified: [list them]. Help me turn the strongest one into a SMART+R challenge statement. Then run it through the Traffic-Light Gate checklist so I can see whether it passes or where it falls short.”

Prompt 3: “I have two challenge statements that passed the gate. Use the Effort vs. Impact Matrix to help me decide which one to pilot first. Then outline what my 7-Day Mini-Pilot would look like for the winner, including what ‘success’ means at the end of day 7.”

• Quantified objectives with baseline and target metrics (not just “improve efficiency”)
• A scored comparison of potential solutions using a credibility scale (0 to 10)
• A compliance check against UAE/KSA regulatory requirements
• A 12-week pilot roadmap with Go, Pivot or No-Go decision points at weeks 4, 8, and 12
• Criteria for stopping the pilot defined in advance (a built-in “kill switch” so you know when to stop, not just when to continue)

Open in NotebookLM

• Start with the evaluation framework (Section 3) when you receive a vendor proposal or identify a tool.
• Distribute sections based on who’s available: technical review, compliance check, and success metrics don’t need to sit with the same person.
• Use the pilot roadmap template (Section 4) to structure a phased plan with clear Go, Pivot, or No-Go checkpoints.
• Return to the framework at increment reviews (weeks 4, 8, 12) to assess whether the solution is delivering what was expected.

This playbook is available on NotebookLM, an AI-powered notebook where you can ask questions in plain language and get answers grounded entirely in the playbook content, nothing drawn from outside it.

The prompts below are designed to take you through a complete evaluation using your own organization’s context. Use whichever is most relevant to where you are in the process.

Prompt 1: “We’re evaluating [name the AI tool] for [describe your use case]. Help me set up an Impact Estimation Table with 3 quantified objectives. I need realistic baseline metrics and target values for a 12-week pilot with a team of [size] and a budget of [amount].”

Prompt 2: “Here’s what we know about this solution so far: [paste vendor claims, demo notes, or reference material]. Score its credibility using the 0-10 scale from Section 3. Where are the evidence gaps? What specific questions should I send back to the vendor before our next meeting?”

Prompt 3: “We’ve decided to proceed with [solution]. Design a 12-week pilot roadmap using the template from Section 4. Include GO/PIVOT/NO-GO criteria at each increment review, the compliance gate requirements for UAE context, and what our kill-switch protocol should look like if results plateau after week 6.”