How I’ve Used AI Over the Past Six Months as CEO

Over the last year, I’ve been working with three to four LLMs consistently: ChatGPT, Gemini, Claude, and Grok, in that order. I’m a pragmatic super-fan of them. I know they hallucinate, gush, and lead the witness, but they also have played a key role in helping me lead and manage a growing, complex organization while cranking out a high volume of work and decisions. 

I wanted to share a few examples of how I leveraged them in my role:

1. Auditing My Time as CEO

In late December, I asked Gemini to audit my Google Calendar. I asked: “Read through my 2025 calendar carefully and tell me what I spent my time on.” Gemini automatically reviewed and categorized everything I did, grouping my activity into themes like board work, staff management, strategy, operations, writing, and external engagement. I then asked AI to assign percentages and calculate the implied dollar value of that time based on my compensation.

Then I asked: “Compare my 2025 calendar to my 2019 calendar and tell me what you notice.” It was revealing, and after that, the questions kept flowing intuitively. 

This simple review exposed places where I was operating at the right altitude and places where I had unintentionally slipped into work that wasn’t, let’s say, the best use of my time. (Although my staff might describe this in spicier terms.)

2. Reviewing Five Years of Board Materials

Like many of you, I submit quarterly CEO commentaries, dashboards, agendas, and other updates prior to scheduled board meetings. After a while, those reports, projections, and comments get a little foggy. You wouldn’t think so, but they do. I asked Gemini to carefully review all my submitted board documents over the past five years and answer some questions:

• What have I promised and not completed? 
• Summarize my CEO commentary themes and tone. 
• Are these commentaries suitable, in your opinion? 
• Review the agendas and tell me if my board would have enjoyed those meetings.
• Grade the board’s agenda focus. Are we governing well based on what you see?

The results were very instructive. I saw plenty of helpful feedback and insights, but also bumped into specific issues that needed to be addressed—things that were likely irritating my board but had been flying under the radar; things I could get ahead of; things I needed to stop; and things I could do to craft board meetings that would be better received. This review would have taken an impossible amount of manual effort without AI. In a matter of an hour, I acquired a compressed line of sight on five years (20 meetings) of board activity.

3. Rewriting and Streamlining Our GUIDE

One of the core resources we share with Charity Partners (“the Guide”) is a 120-page online resource that has grown and grown (and grown) over the years. Some might call it comprehensive, but I had a feeling it was actually becoming bloated. Over time, sections had been layered on top of one another. It was thorough, but it wasn’t tight. So here’s what I did in support of our newer U.S. expansion.

I made a PDF of the whole thing, uploaded it to ChatGPT, and asked it: “Please slowly, carefully read the following Guide and grade it for me. Is it clear? Is it easy to read? What are the different voices? Where is it strong, weak, too long, or too short. I then provided our founding documents—vision, mission, and values—and asked it to rewrite the Guide in the same voice and tense while reducing the word count by at least thirty percent.

I remained fully engaged in the editing process, refining where needed and protecting the intent of every section (most of which I’d written 15 years earlier) that AI might not fully understand.

Within a week we reduced the document by 26,000 words (a 32 percent reduction) and streamlined it from nine sections to six. The final version was clearer and easier to use operationally. What might have taken months of committee work was completed in days without sacrificing alignment, and it was now imminently repeatable.

4. Analyzing Campaign Data

I’ve also experimented at the operational level. Every week during one of our national peer-to-peer events, I export campaign data by location and compare its performance week over week, month to date, and campaign to date — comparing it against last year’s results and setting goals.

Instead of manually scanning spreadsheets, I’ve now been asking AI to surface what matters:

• Which regions are underperforming relative to last year? 
• Where is campaign growth concentrated? 
• Does that growth correlate with staffing changes or onboarding timelines?

Now, I’m able to interpret patterns instead of just looking at raw numbers. That movement from data to insight happens faster, and it sharpens the conversations I have with the team. I’ve also used this review to build out predictive algorithms to use in our tech stack. 

5. Pressure Testing Organizational Design

As we separated Canadian and U.S. operations and clarified ownership, I used all four AIs to stress test structural decisions. I explored what typically breaks when ownership is split across borders, where decision rights blur, and which risks emerge when a CEO temporarily leads a new regional team. I also asked what clarity must exist before hiring a Director of U.S. Operations and what common failure modes appear in nonprofit expansion.

Now, to be clear: AI did not dictate the structure. I used it as a thought partner, and it helped me explore ideas, decisions, and potential upstream consequences. It helped me surface ambiguity before that ambiguity became conflict, which in leadership is often half the battle.

6. Reflecting on Leadership

I often bring real scenarios into the conversation, such as conflict, uncertainty around strategic timing, or questions in difficult decisions. The system can pull context from prior discussions, identify recurring themes in my leadership style, and challenge assumptions I may not see clearly. The key is staying in control of the dialogue. I direct the conversation deliberately and avoid letting it steer me prematurely toward conclusions.

If you are not using AI regularly, this kind of feedback will not be as helpful. But if you are using it actively, it develops a line of sight into patterns and details you may have forgotten. Its memory is often better than mine, which makes its reflections more precise and grounded.

Because I use it consistently, it accumulates context across months of prompts, conversations, and reflection. That longitudinal perspective allows it to notice drift in my priorities, surface contradictions, and remind me of earlier commitments.

It does not replace discernment. The authority remains with me. But the feedback loop becomes faster, broader, and more informed.

7. Moving from Blank Page to Concept

When developing new event concepts or refining existing ones, AI is particularly useful in the early stages. If I have a partially developed idea, I outline it and ask for critique:

• Is this positioning aligned with our operational reality, or are we promising more than we can deliver?
• Where might smaller charities encounter friction?
• How might different demographics interpret the name?

It doesn’t replace human brainstorming, but instead helps accelerate the transition from blank page to structured concept. It also exposes weak logic quickly and forces coherence, which helps when it’s time to share an idea more broadly. It’s also really helpful to let AI surface first principles that should be influencing innovation.

8. Strengthening Writing and Thought Development

I write frequently, whether for board updates, internal strategy memos, or public reflection (ahem, like this article). I’ll often include AI in my process now, uploading a strong but imperfect draft and asking it to challenge my assumptions and surface what a skeptical reader might question.

The ideas remain mine. The judgment remains mine. But the time required to pressure-test those ideas has decreased significantly. Work that once consumed a full day can now be completed in a fraction of that time, often with greater rigour.

A Simple Prompt to Try

If you’re experimenting with AI in your own leadership, it may be worth stepping back and asking a broader question: where is this actually strengthening your judgment, and where is it just convenient?

One practical way to test that is to ask it directly:

• Based on our conversations, how am I using AI most often?
• Where have I seen the highest return on time invested?
• Where am I underusing it?
• What patterns do you see in my questions?
• If I wanted to double the return on how I use AI, what should I focus on?

The hard thinking still belongs to you. But if AI helps you see more clearly and act more deliberately, that’s leverage. Often, the reflection itself is the real return.