Cherie Blair CBE KC joined TEAM LEWIS for a keynote on why decision-making can feel harder than it should in a world drowning in information. She opened with a question that made the room go quiet: think about your last high‑stakes call – did the data help you move, or did it give you a reason to wait?
Her message was simple. Evidence matters, but it has limits. Data can tell you what’s happened, not always what’s possible. And now that AI is part of day‑to‑day work, confident outputs are easier than ever to mistake for correct ones. In the end, leaders still have to judge, decide and own the outcome.
Drawing on examples from both her legal career and the work of the Cherie Blair Foundation for Women, which she founded, she illustrated what leadership looks like in practice: making decisions when the information is incomplete and taking responsibility for them. She also highlighted the Foundation’s latest research into women entrepreneurs and AI, published in partnership with Intuit and TEAM LEWIS.
While the headline finding showed a dramatic increase in AI adoption among women entrepreneurs, the deeper analysis revealed a more nuanced picture. Women who embed AI into core business functions such as finance, operations and strategic planning are seeing significantly greater business benefits than those using it primarily for everyday administrative tasks.
It was a powerful reminder that headline figures rarely tell the whole story – and that good leadership means looking beyond the data to understand what is really happening.
If you work in marketing (like me), you know the drill: dashboards, weekly reports, performance insights. The data isn’t the hard part. The hard part is calling it – and moving.
Here’s what I took away – and what leaders can do next.
1) The new procrastination: “let’s just check one more thing”
We’ve turned hesitation into a process.
A decision comes up. The stakes are real. And instead of deciding, we ask for another breakdown, another validation, another round of “just to be safe”.
Sometimes that’s responsible. Sometimes it’s just delay with nicer formatting.
If your meetings sound like:
- “Can we look at it another way?”
- “Let’s sanity-check the model.”
- “Let’s wait for next month’s numbers.”
- “We need alignment first.”
… you’re not alone. The question is whether those checks are getting you closer to a call – or keeping you comfortably busy.
2) The messy truth: more metrics don’t make it clearer
More information rarely creates clarity. Most of the time it creates debate.
The fastest teams aren’t tracking less because they don’t care. They’re tracking less because they know what matters.
A simple standard: if a metric can’t answer “so what?” or “what changes because of this?”, it doesn’t belong in the headline story. Keep it in the background, but don’t let it run the meeting.
3) The uncomfortable test: are we learning, or stalling?
Sometimes “we need more evidence” is true. But sometimes it’s a polite way of saying: “I’m not ready to own the decision.”
A quick gut-check: is this helping us decide – or helping us avoid deciding?
If it’s avoidance, you need three things quickly:
- A decision date
- A named owner
- Agreement on what “good enough” evidence looks like
Without those, you don’t have a decision process. You have a discussion habit.
4) Data doesn’t speak for itself. People do
The most persuasive chart in the room can still be incomplete – not because anyone is misleading you, but because measurement always reflects choices.
So ask:
- What’s missing?
- What did we choose not to measure, and why?
- Whose definition of success are we using?
- What changes when you look by market, audience or segment?
For international teams, this is where good decisions are won or lost. Global averages can hide local reality. And local reality is where outcomes happen.
5) AI is fast. That doesn’t make it right
AI is useful. It’s increasingly standard. It also has a talent for sounding certain.
That’s the risk: an answer that sounds certain can shut down the questions we still need to ask.
The fix isn’t to avoid AI. It’s to use it with guardrails:
- Treat outputs as a strong first draft, not the final word
- Check what it’s based on (inputs matter)
- Challenge it like you would any recommendation
- Keep a named person accountable for the decision
AI can speed things up. It can’t carry consequences.
6) “Everyone’s using it” is not a success metric
Saying “we’re using AI” is the start of the story, not the outcome.
The questions that matter are blunt:
- Is it improving decisions – or just speeding up admin?
- Are we getting better outcomes – or just more output?
- Are we building capability across the team – or leaving it to the same few experts?
Adoption tells you it’s being used. Impact tells you it’s working.
7) Someone still has to decide
There’s always a moment where the slides end, the discussion goes round in circles and someone has to own the decision.
That’s leadership:
- Choosing without perfect certainty
- Moving without total agreement
- Explaining the decision clearly
- Owning what happens next
No dashboard does that. No model. No AI tool.
If you’re feeling blinded by data, you don’t need less information. You need sharper judgment and the confidence to move while things are still messy.
Data helps. AI helps. But responsibility doesn’t shift. The decision still sits with us.
Find out how TEAM LEWIS can help you move from data to decisions.
