Make AI Your Strategic Thought Partner: Here’s How

Artificial intelligence is stepping into strategic roles in some nonprofit executive suites. 

Among leaders who use AI, the share who report applying it to “organizational strategy” jumped from 33 percent in 2024 to 65 percent in 2025, according to surveys by the consulting firm BDO. 

But that doesn’t mean the work is deep or widespread yet. “There’s an adoption gap and an interest gap in that next-level usage, which is the strategic piece,” says Emily Yu, chief partnerships and program officer at Newman’s Own Foundation, which makes grants to improve children’s lives. “But I think it’s actually a healthy tension” — one rooted in appropriate caution about ethics, data privacy, and the trustworthiness of AI outputs.

Done well, though, the payoff is real. “Not only can it lessen the number of tasks, but it can also increase the space for strategy,” says Mike Mitchell, managing partner of Nonprophet AI, a consulting firm in Baltimore. “And if you increase the space for strategy, you create an opportunity for leveraging the talent that you have.”

Start with a high-stakes strategic question.

One of Mitchell’s clients, the Karsh Social Service Center in Los Angeles, has used AI to refine its “theory of change” — a road map for how the nonprofit’s work will produce the change it wants to see.

The Karsh Center provides health care, legal aid, food assistance, and mental-health services to underserved populations who live in Los Angeles’s Koreatown. Lila Guirguis, its executive director, says she wants a good theory of change because she seeks to move the organization “from just counting inputs like how many people show up to a class, or how many people show up to the food pantry, to really looking at impact data.” To improve the center’s theory of change, Guirguis used AI to come up with a new answer to the question: “Why does the Karsh Center exist?”

AI works best when users employ precise vocabulary and are very descriptive about their needs.

Guirguis used a strategic-planning coach within an AI “command center” that Mitchell and she developed together, custom-built using the large language models best suited for each module. She wrote a prompt for the strategic-planning virtual coach, saying she wanted a theory of change that would be in harmony with the center’s strategic plan and would follow principles laid out by Simon Sinek, a leadership and communications consultant. 

She asked for three options for answers. That prompt began a conversation between the AI coach and Guirguis, with her asking for adjustments in language, in part to avoid cliches. The result: “We believe every person — regardless of background, language, or circumstance — deserves access to the support and services needed to live a life with dignity.” Guirguis says she now believes the organization’s theory of change and strategic plan are tightly integrated.

Lead with strategy, not tools.

Compass Pro Bono also uses AI as a strategic-thought partner as it works to help other nonprofits find skilled volunteers and navigate a wide range of issues, including using AI. Now with offices in five cities, Compass Pro Bono uses AI to help it find organizations with similar missions as it expands nationally, seeks out corporate partners that can sustain its expansion, and learns how it could help local nonprofits in target cities. 

When it comes to program strategy, Remy Reya, Compass Pro Bono’s director of AI and thought leadership, says AI “is never the first or last stop in our choices, but it has gotten incredibly good at providing thoughtful, targeted guidance.” 

Nonprofits should first identify workflows where AI could be a thought partner, says Daria Seriakova, senior director of technology and operations at Common Impact. She encourages nonprofits “to take the bird’s-eye view of implementation and lead with strategy rather than tools.”

Shape your AI into the collaborator you need.

When working with AI, Reya says organized workspaces — called “projects” in Claude and ChatGPT, and “notebooks” in Gemini — are where you start to get “exponential value” from AI tools. 

The feature allows users to load up context, background, and knowledge so that every time a user interacts with a project, the AI agent draws on that library of information. 

Making the switch from one-off prompts helps you build valuable context that the platform remembers. Instead of getting the basic, “How can I help you today?” prompt, users can get an executive briefing or a strategic-plan monitoring dashboard. 

Write a job description and give it to your AI assistant.

When nonprofits work with AI on strategic projects, Reya says, they aren’t hiring a partner — they are building one. AI can be obsequious, he and others say, and nonprofit executives may need to tell AI assistants that they want an intellectual sparring partner, not a sycophant. If executive directors would like their AI assistant to act like a skeptical board member, they can ask for that. In short, he says, write a job description and give it to your AI assistant. 

Executives who are good people managers should also excel at shaping AI partners to be productive. Onboarding AI can be similar to managing an intern, says Reya. He wouldn’t ask an intern to write an annual report without supplying previous reports and relevant data. And if the first draft was poor, “I wouldn’t just say, well, that’s just how interns are, and then go and do it myself. I would give them feedback and iterate and nudge them toward the right output.”

In working with AI, Reya says, “I encourage people to get in the mindset of, ‘If I were working with a human, what would they need to know to do this well?’”

Write prompts like you’re briefing a new hire.

AI works best, says Mitchell, when users employ precise vocabulary and are very descriptive about their needs. He encourages clients to write prompts that might be pages of instructions, not just a few sentences. And users can apply AI to improve prompts.

A nonprofit executive might ask an AI assistant to serve as a financial analyst, Mitchell says, comparing quarterly financials to strategic-plan priorities, determine where spending isn’t aligned with priorities, and then present findings in a variety of formats. The nonprofit could input the prompt and ask the AI assistant for improvements. An AI assistant might respond by asking for guidelines in correctly interpreting the data, prioritizing recommendations, and seeking to know who the audience is, so it can tailor the presentations to them. 

Don’t outsource critical thinking.

Yu doesn’t discount the idea of using AI in strategic planning but says discussion and disagreement among staff members, leadership, and a board should not be eliminated. If stakeholders are not heavily involved in drafting a plan, she says, “then we inherently lose the core value of what a strategic plan is meant to do.”

Experts advise resisting the temptation to use AI writing, no matter how polished it might seem, and rewrite it to make sure it has the voice of the organization or the author.

Use the time AI saves to check facts, sources, and validate perspectives, says Reya, who gives a presentation on “keeping your brain in the driver’s seat” while working with AI. He encourages users to tell an AI assistant explicitly what its role should be. Reya might instruct an AI tool that serving as his thought partner “means my voice, ideas, and critical thinking must stay front-and-center throughout all of our collaborations. Your job is to augment my cognition and creativity.”