For CEOs and Founders

How CEOs Can Use AI Without Losing Their Edge

You can delegate analysis to Claude or ChatGPT, but if you let those tools shape the question before you see it, you stop thinking like a CEO. Your board expects your perspective, not a summarised version of what an AI thinks. The risk is real: you lose the instinct that caught problems early when you were smaller.

These are suggestions. Your situation will differ. Use what is useful.

Download printable PDF

Separate analysis from framing

When you ask an AI tool to analyse a market opportunity or competitive threat, it will give you a shaped answer based on how you framed the question. Your job is to challenge the frame before you accept the analysis. Read the raw data or talk to your operations lead first. Then use the AI to test your own hypothesis, not to replace it. This keeps you from inheriting someone else's assumptions as your own strategic view.

›Before running a board situation through Claude, write down your own reading of it in one paragraph
›Ask your CFO or COO for their instinctive take on a problem before you feed it to an AI tool
›When an AI summary feels too clean or too certain, go back to the original sources it drew from

Keep board communication personal and contrarian

Your board does not want a polished AI summary of the quarter. They want to know what you think is actually happening and what keeps you awake. If you send them a board deck that reads like it was synthesised from market reports, you lose credibility and lose the chance to test your judgment against experienced advisors. Write your own strategic narrative first. Use AI to help you stress-test it or find data gaps, not to generate it.

›Draft your board letter yourself before you open any AI tool
›Ask one board member directly: 'What do you need to hear from me that you can't get from the numbers?'
›When you use AI to fill in market context, mark it clearly so the board knows where your thinking ends and data synthesis begins

Practise the skill you most need to keep

The pattern recognition you built over years of operational experience is worth more than any AI analysis. But it atrophies fast if you stop using it. When you delegate too much analytical work to your tools, you stop noticing the small signals that used to alert you early. Spend time each week on the raw business data yourself: customer calls, pipeline conversations, operational metrics. Use that directly observed context to question what your AI tools tell you.

›Sit in on one customer call each week without preparing talking points beforehand
›Review your top 20 pipeline deals yourself instead of reading a digest
›Keep a notebook of what you notice that does not appear in reports, then ask why it is missing

Create a rule for when you do not use AI

Your strategic instinct works because it absorbs information across multiple dimensions at once. If you offload all your thinking to AI, you stop building that muscle. Create a clear rule: certain decisions or communications you will always do without AI first. For many CEOs, this is board communication, investor updates, and big strategic pivots. For others, it is the weekly operational review or the quarterly strategy session. The point is to protect the decision-making space where your judgment matters most.

›Decide which three types of communication you will write yourself before using any AI help
›Tell your executive team which decisions you want their thinking on before you synthesise it
›Review the decisions you made without AI help and the ones you delegated; notice which ones you owned more fully

Build a contrarian into your decision process

AI tools are consensus machines. They synthesise what has been written and said before. The problems that catch you off guard are usually things the AI has not seen in its training data. Find one person on your leadership team who naturally disagrees with consensus. Make sure they see your AI summaries and ask them specifically where the tool is probably wrong. This person is protecting you from the biggest risk of AI adoption: the quiet drift toward groupthink that feels like clear analysis.

›Ask your most contrarian leader to review AI summaries with one question: 'What is this missing?'
›Track the times someone on your team spotted a risk that your AI analysis did not flag, then build that into your next brief
›When your board raises a concern, check whether any of your AI tools flagged it first; if not, that is a gap to understand

Key principles

1.Use AI to expand your access to information, not to replace your thinking before you engage with a problem.
2.Your strategic instinct is built on pattern recognition from direct experience; protect the time you spend observing your business unmediated.
3.Board and investor communication should reflect your own judgment first, with AI as a secondary tool for stress-testing and data filling.
4.The contrarian voice on your team is your insurance policy against the slow drift toward consensus that AI tools can reinforce.
5.The decisions where you do not use AI are the ones where you stay sharp; choose them deliberately.

Key reminders

Write your own take on a strategic issue in one paragraph before you ask ChatGPT to analyse it; compare what you missed to what the tool found
When your CFO or COO brings you an AI-generated insight, ask what they think independently first
Create a monthly review where you look at decisions made with heavy AI input versus light AI input; notice which ones you understand more deeply
Set a rule that all board materials include one section written entirely by you without AI assistance
Keep one hour each week for operational observation that has no AI synthesis: customer calls, site visits, unstructured conversation with frontline team members