Strategic philanthropy learning agenda builder · free AI skill

Turns a foundation's strategic priorities into a structured learning agenda with prioritized learning questions, evidence requirements, matched methods, a phased timeline, and explicit decision links showing how…

Purpose

Convert a foundation's strategy into a decision-oriented learning agenda — a structured set of learning questions, each mapped to the evidence needed, the method to gather it, a timeline, and the specific grantmaking decision it will inform. Solves the common problem of learning activities that generate reports nobody uses because they were never tied to a decision.

When to Use

Invoke this skill when the user asks to:

Build or refresh a learning agenda for a strategy, initiative, or portfolio
Design an evaluation or learning plan tied to strategic priorities
Turn a foundation strategy or theory of change into learning questions
Connect evaluation activities to specific grantmaking decisions
Prioritize which learning questions to invest in with limited evaluation budget
Prepare a learning agenda to socialize with program staff, leadership, or the board

Inputs Required

Before starting, collect:

Strategy area or portfolio: The initiative, program area, or strategy the learning agenda serves
Strategic priorities or goals: What the foundation is trying to change or achieve
Theory of change or key assumptions: The causal logic and the assumptions most in doubt (if available)
Decisions on the horizon: Upcoming grantmaking or strategy decisions this learning should inform (e.g., renewal, pivot, scale, sunset)
Time horizon: Planning period for the agenda (e.g., 12 months, 3 years)
Evaluation capacity: Internal staff, external evaluators, budget range, or data infrastructure available
Audience for findings: Who will use the learning — program officers, leadership, board, grantees, field

If theory of change or capacity details are missing, infer reasonable assumptions from the strategy and note them explicitly.

Instructions

Step 1: Surface Candidate Learning Questions

Derive learning questions from three sources, working across all of them:

Theory of change assumptions: For each major causal link, ask "what would we need to see to believe this is working?" The riskiest, least-evidenced assumptions become the strongest question candidates.
Upcoming decisions: For each decision on the horizon, ask "what do we need to know to make this well?" Reverse-engineer questions from the decision.
Strategic uncertainties: Where is leadership genuinely unsure — about approach, context, grantee capacity, or field readiness?

Organize candidate questions into four types:

Implementation questions: Are grants and activities happening as intended?
Effectiveness questions: Are activities producing the intended outcomes?
Contribution questions: How much is the foundation's role driving observed change vs. other factors?
Strategic questions: Should the strategy itself change — approach, allocation, or scope?

Aim for 12–18 candidate questions before prioritizing.

Step 2: Prioritize the Question Set

Score each candidate on three dimensions (High / Medium / Low):

Decision relevance: Is a real decision waiting on this answer?
Uncertainty: How much genuine doubt exists? (Questions we already know the answer to are not worth funding.)
Feasibility: Can we credibly answer it given capacity and timeline?

Select 5–8 priority questions. Favor questions that are high on decision relevance AND high on uncertainty. Explicitly park lower-priority questions in a "not now" list so they aren't lost.

Step 3: Specify Evidence Needed Per Question

For each priority question, define:

What evidence would answer it: The specific data, observations, or perspectives required
Evidence standard: How rigorous the evidence needs to be to support the decision — a renewal at scale demands more than an early directional read
Whose perspective matters: Grantees, participants, community members, field experts, staff
What "good enough" looks like: The threshold that lets the decision-maker act with confidence

Step 4: Match Methods to Questions

Assign a method appropriate to each question type and evidence standard. Do not over-engineer.

Implementation questions: Grantee reporting, site visits, dashboard/output tracking, structured check-ins
Effectiveness questions: Outcome tracking, pre/post measurement, comparison or benchmark analysis, developmental evaluation
Contribution questions: Contribution analysis, process tracing, key informant interviews, outcome harvesting
Strategic questions: Landscape scans, expert convenings, portfolio reviews, sensemaking sessions

For each, note the method, the likely data source, and whether it runs internally or needs an external evaluator.

Step 5: Sequence the Timeline

Lay out a phased timeline across the planning horizon:

Phase 1 — Foundation-setting: Baseline data, method design, evaluator selection
Phase 2 — Active learning: Data collection and interim reads on time-sensitive questions
Phase 3 — Synthesis and decision: Sensemaking, findings integration, and delivery timed to arrive before each linked decision
Anchor each question's key milestones to the decision it informs — evidence that arrives after the decision is wasted.

Step 6: Link Findings to Grantmaking Decisions

For each priority question, write a one-line decision link: "This finding will inform [specific decision] by [date] by helping us [judgment to be made]." Define a lightweight sensemaking cadence — who reviews findings, how often, and how conclusions get translated into portfolio action. Close the loop: name where learning will re-enter the strategy.

Output Format

Structure: A prioritized learning agenda table (one row per priority question with question, type, evidence needed, method, timeline, linked decision), followed by supporting narrative sections
Length: 900–1,400 words
Format: Structured document with a central table plus prose sections; parked questions as a short list
Sections: Learning agenda overview → Prioritized question table → Prioritization rationale → Methods and capacity summary → Phased timeline → Decision links and sensemaking cadence → Parked ("not now") questions
Tone: Evaluation officer — rigorous, decision-focused, plain-language; skeptical of learning-for-its-own-sake

Quality Checklist

Before delivering output, verify:

[ ] The agenda contains 5–8 priority questions, each tagged by type (implementation / effectiveness / contribution / strategic)
[ ] Every priority question has a named method matched to its evidence standard — not a default "we'll evaluate it"
[ ] Every priority question has an explicit link to a specific grantmaking decision and date
[ ] Prioritization rationale references decision relevance and uncertainty, not just interest
[ ] Timeline milestones land before the decisions they inform, not after
[ ] A "not now" parked-question list is included so deprioritized questions are preserved

Notes

The most common failure mode is a learning agenda with too many questions and no decisions attached — a tight agenda of 5 decision-linked questions beats a comprehensive one of 20.
Match method rigor to the stakes of the decision; early directional questions do not need randomized designs, and high-stakes scale decisions should not rely on anecdote.
Build in sensemaking as a scheduled activity, not an afterthought — findings that are collected but never collectively interpreted rarely change grantmaking.
Revisit the agenda when strategy shifts; a learning agenda is a living document, not a one-time deliverable.

How to use this skill

This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.

Installing in Claude Code:

~/.claude/skills/<folder-name>/SKILL.md — auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)
<your-project>/.claude/skills/<folder-name>/SKILL.md — auto-loads only when working in that project (recommended for project-specific skills)

Restart Claude Code or open a new session to register.

Using in Claude.ai, ChatGPT, Gemini, or other AI tools:

Drag this file into a new chat, or copy and paste its contents
Send your task or request as your next message — the AI will apply the skill where it fits

For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills