Future:Fit Leadership — Blog 5

The Augmented Decision Room

AI is already entering the executive decision room.

Sometimes openly.

A leader asks an approved AI assistant to summarize market evidence, model scenarios or challenge a proposal.

Sometimes invisibly.

An executive arrives with an AI-generated briefing. A team member uses AI to prepare the business case. A consultant’s recommendation has been shaped by machine-generated analysis that nobody in the room can see.

Either way, AI is becoming part of organizational decision-making.

The question is no longer whether it is in the room.

The question is whether leaders know what role it is playing.

Adding AI does not automatically improve a decision

It is tempting to imagine AI as another expert at the table: faster, more analytical and less influenced by organizational politics.

But AI is not a neutral participant.

Its outputs reflect the data, design and instructions behind them. It can reproduce bias, omit context and express uncertainty with the same fluency as confidence.

It can also amplify the assumptions of the person operating it.

Ask AI to support a preferred strategy and it will often build a compelling case. Ask it to dismantle the same strategy and it may produce an equally persuasive argument.

That does not make AI useless.

It makes leadership discipline essential.

The NIST AI Risk Management Framework emphasizes that trustworthy AI requires ongoing governance, mapping, measurement and management—not a one-time technical check.

It also stresses the importance of clearly defining human roles and responsibilities when AI contributes to decisions.

That is the heart of the augmented decision room:

AI may inform the decision. Humans must remain accountable for it.

AI can widen perspective—or standardize it

One of AI’s greatest strengths is its ability to expand the range of possibilities a team considers.

It can identify missing evidence, simulate stakeholder perspectives, generate alternative scenarios and act as a tireless red team.

But there is another possibility.

If every organization uses similar systems with similar prompts to solve similar problems, strategic thinking may begin to converge around the most statistically probable answer.

Recent Harvard Business School analysis warns of an emerging “AI groupthink” problem: standard AI use can push organizations towards increasingly similar, average ideas.

In other words, AI can challenge groupthink inside one room while creating conformity across an entire market.

The advantage will not come from having access to AI.

Most competitors will have access.

It will come from the quality of the questions, perspectives and judgment surrounding it.

The Augmented Decision Protocol

Before using this protocol, decide whether AI should be involved at all. Use only approved, secure systems, and never enter confidential, personal or commercially sensitive information without authorization.

For an important decision, move through six stages.

1. Frame before prompting

Define the decision before asking AI for an answer.

What are we deciding? Why now? Who owns the final call? What constraints and values must shape it?

Write the decision question in one sentence.

If the team cannot agree on the question, it is not ready to evaluate answers.

2. Establish the human view

Ask decision-makers to record their initial position independently before seeing AI-generated analysis.

What do they believe? What evidence supports that belief? How confident are they?

This step matters because an early AI recommendation can anchor the entire discussion. Capturing the human view first makes shifts in thinking visible.

The objective is not to privilege intuition.

It is to prevent machine output from becoming the room’s unexamined starting point.

3. Expand with AI

Use AI to widen the field rather than select the winner.

Ask it to:

• Generate several genuinely different options
• Identify missing information
• Surface assumptions behind each option
• Model best-, expected- and worst-case scenarios
• Adopt the perspectives of customers, employees, regulators and competitors
• Highlight where its analysis is uncertain

Do not settle for one confident response.

Use different prompts, approaches or approved models where appropriate. Compare the results and investigate important disagreements.

4. Red-team the preferred option

Once a direction begins to emerge, turn AI against it.

Ask:

• Why might this decision fail?
• What would a sceptical expert challenge?
• Which data could contradict our conclusion?
• What second-order consequences might appear?
• Who may be unintentionally disadvantaged?
• What are we not seeing because of how we framed the problem?

Then assign a human challenger.

AI can generate objections, but a person must test whether they matter in the organization’s real context.

5. Decide and document

The accountable leader makes the decision.

Record:

• What was decided
• What evidence was used
• How AI contributed
• Which assumptions remain uncertain
• Who challenged the recommendation
• Who owns the outcome
• What would trigger reconsideration

Documentation should not become bureaucracy.

It should make responsibility visible.

If nobody can explain how the decision was reached, the organization cannot learn from it later.

6. Review the outcome

Return to the decision after an agreed period.

Which assumptions proved correct? What did the humans notice that AI missed? What did AI surface that the team had overlooked? Did the process improve the result—or merely accelerate it?

This final step turns individual decisions into institutional learning.

Without review, organizations repeat mistakes with greater efficiency.

Better data does not remove politics

Some leaders hope that AI will make executive decisions more objective.

But important decisions involve competing priorities, unequal consequences and different interpretations of what “better” means.

Data can clarify trade-offs.

It cannot eliminate them.

A model may identify the most efficient restructuring option. It cannot decide whether the human cost is consistent with the organization’s values.

It may rank candidates according to defined criteria. It cannot determine whether those criteria reflect historical bias.

It may calculate expected return. It cannot decide what level of risk is morally acceptable.

These are not weaknesses in the technology.

They are reminders of the leader’s responsibility.

The decision room must remain human

An augmented decision room is not one in which AI speaks most loudly.

It is one in which the combined process produces better questions, broader evidence, more credible challenge and clearer accountability.

AI should make weak assumptions harder to hide.

It should make dissent easier to explore.

It should help leaders see beyond the limits of their own experience.

But it must never provide a convenient place to transfer responsibility.

Your Future:Fit challenge: Choose one consequential decision this month and use the six-stage Augmented Decision Protocol. At the end, ask the team: Did AI deepen our judgment—or simply increase our confidence?

The future of decision-making is not artificial.

It is augmented, accountable and unmistakably human.

#FutureFitLeadership #AugmentedLeadership #DecisionMaking #AIGovernance #ArtificialIntelligence