Leadership teams have more data than ever, and now AI to analyse it. Yet decisions have not improved at the same pace.
Deloitte's 2026 Global Human Capital Trends recommends treating decision-making as a strategic discipline. That starts with knowing the traps.
The six traps
1. Skipping the problem
Teams jump to solutions before agreeing what problem they are solving. Different people end up solving different problems.
Counter it: Write the problem in one sentence before discussing options. Agree it first.
2. One process for everything
Teams use the same approach for small, reversible calls and big, irreversible ones. Small decisions get slowed down. Big ones get rushed.
Counter it: Sort decisions by how reversible they are. Move fast on the reversible ones.
3. Anchoring on the first option
The first workable idea becomes the reference point. Kahneman's research shows how strongly early anchors shape later judgement. AI drafts can become that anchor.
Counter it: Generate at least three options before evaluating any. Form your own view before reading the AI's.
4. Mistaking agreement for quality
Quick agreement feels like a good sign. Often it means people did not feel safe to disagree. Edmondson's research links psychological safety to speaking up.
Counter it: Ask someone to argue the opposite case. Ask quieter members first.
5. Treating data as a substitute for judgement
Data and AI output can feel objective. Every analysis still rests on choices about what to measure and include.
Counter it: Ask what the data cannot tell you. Ask who checked the AI's answer, and how.
6. Sunk cost
Past investment keeps failing projects alive. McKinsey's State of Organizations 2026 found 43% of leaders say they divested assets too late, or failed to.
Counter it: Ask: if we were starting today, would we choose this? Set stop rules at the outset.
Build this capability in your organisation
Growth Performance designs bespoke programmes that turn insight into measurable behaviour change.
The common thread
Each trap comes from the conditions around a decision. Lovallo and Sibony found that a good decision process mattered more than analysis in explaining results.
The conditions are clear problem framing, safety to disagree and the discipline to check. AI does not remove the traps. It can make them faster.
What this means for leaders
Checking AI-supported decisions properly is part of Adaptation: people directing, combining and checking AI in real work.
Try this
Review your team's last three big decisions. Which traps appeared? Pick one counter-measure and use it at your next meeting.
The Team Decision-Making Guide helps you diagnose your team's patterns.