Before asking which AI tools to use, ask where AI actually belongs.
With more and more AI tools available, the question many organizational leaders are asking themselves is: “Which tools should we use?” However, a more productive place to start is, “Where does AI actually belong in our work?”
People often cite speed and scale as AI’s greatest strengths. But output quality directly ties to input quality, and volume alone doesn’t produce insight. The organizations getting the most out of these tools aren’t necessarily the ones that move the fastest, they’re the most intentional. For purpose-driven organizations, people come first, and centering the people an organization serves, along with the people delivering on the mission, is a critical first step toward understanding when and how to use AI.
Thinking About Where AI Fits
A recent framework I learned in IDEO’s Human-Centered Research with AI webinar introduced a practical lens for using AI. The following 2×2 matrix uses two axes (Accelerate vs. Deliberate & Expand vs. Interrogate), producing four distinct modes. Before stepping into a task, this is a way to categorize it and help you understand what you’re actually asking AI to do for you.
Taking each quadrant a step further, I’ve outlined a handful of examples that speak to common problems you may seek to solve in your work.
“I know what I need, but don’t have the capacity to get there.”
Widen the aperture (Accelerate + Expand) | AI expands your capacity to take in information and generate foundational content quickly. Best used for divergent points of your process or when dealing with high volume:
- Compile a landscape scan of peer organizations before a board retreat.
- Generate a first draft of a grant narrative from an RFP and existing program materials.
- Synthesize participant feedback from a program survey into key themes before a staff planning retreat.
“I have information, but I might not be seeing the full picture.”
Notice nuance (Expand + Deliberate) | AI surfaces patterns and alternative perspectives before synthesis. Human thinking stays in the lead with these examples in identifying perspectives to explore and making decisions with those perspectives, but AI helps broaden the view:
- Read a set of participant testimonials and surface themes that don’t appear in the formal outcome reporting.
- Review your last three annual reports and identify what story they’re collectively telling, then assess whether that matches your intended narrative.
- Review a job description and identify which kind of candidate it’s actually likely to attract, versus the kind of candidate you say you want.
“I’ve defined the work, but want to challenge my thinking.”
Challenge assumptions (Interrogate + Deliberate) | AI as a thought partner to critique ideas, check for bias, and provoke divergent thinking.
- Feed a program model or theory of change into AI and ask it to identify unstated assumptions or gaps in the logic.
- Share a fundraising case for support with AI and ask it to argue against it.
- Upload your mission statement and current program portfolio into AI and identify any drift between what you say you do and what you actually do.
“I have data, but I’m not sure what it’s really telling me.”
Immerse in data and context (Accelerate + Interrogate) | AI plays out hypotheses and surfaces gaps quickly.
- Review your program outcomes data and identify which results you’re reporting and which you’re leaving out.
- Revisit a year’s worth of program intake data to identify what the organization doesn’t yet know about its participants.
- Share a funding report or funder update with AI and ask what questions the data raises that the narrative doesn’t address.
Knowing where AI fits is only half the equation. Getting a useful output depends on how well you ask the question.
Getting Prompts Right
We’ve outlined several use cases and mindsets appropriate for AI use and described the work at a high level. But you’ve just opened Claude, and you’re greeted with the question, ‘How can I help you today?’ Where do you go next?
Generative AI works best when it knows what role it’s playing and how to play it. COSTAR is a prompting framework with six elements (Context, Objective, Style, Tone, Audience, Response) that helps you structure your prompts for AI so the output is actually useful. Consider the following prompt you could use when talking to AI:
| Context: You are a senior leader preparing all-staff communication to introduce your organization’s newly completed strategic plan. Staff have been engaged throughout the planning process, but they haven’t seen the final framework. The process took longer than expected, and there were moments of uncertainty along the way. |
| Objective: Draft a message to staff that introduces the final strategic plan, acknowledges the work it took to complete it, and sets a clear and optimistic tone for implementing the plan. |
| Style: Personal and direct, written in the senior leader’s voice, not overly formal [see attached examples of historical writing samples]. |
| Tone: Celebratory but honest. Forward-looking while acknowledging the team effort it took to get here. |
| Audience: A full staff team, with some closely involved in the planning process and others who are hearing the details for the first time. |
| Response: One page, two to three paragraphs, no headers, with bracketed placeholders for the organization name and where a link to the full plan document would be inserted. |
The more specific the prompt, the more useful the output. COSTAR works because it forces clarity from the very beginning and aligns with the same approach that makes any planning process effective. Attaching reference documents (e.g., past reports, existing materials, or examples of your own writing) can further strengthen your results.
Maintaining Human Centeredness
No matter how you use AI, you should start and end your work with human insights. Start with the questions only your team could ask, informed by the context of those you serve, and your lived experience in the work. Let AI do the work in the middle: synthesizing, reviewing drafts, highlighting patterns, and challenging assumptions. Then bring human judgment back in to review, revise, and ensure the work is still your own. Think about AI as a tool that protects the humanity of your work, freeing up capacity for the conversations, relationships, and judgment that only people can provide.
AI is only as purposeful as the thinking behind it. Figuring out where AI fits is a strategic question, and it deserves the same intentional thinking you’d bring to any organizational decision. If you’re ready to think that through, we’d love to help.
