Most organisations preparing for AI ask broadly the same questions:
- Is our data ready?
- Do we have the right technology?
- What are the security and governance implications?
- Do our people have the skills to use it?
These are all sensible questions, but there is another form of readiness that receives far less attention.
Is the organisation actually capable of acting on what AI tells it?
An organisation can have excellent data, sophisticated technology, robust governance and capable people but still struggle to extract meaningful value from AI.
The problem may not be AI readiness at all - it may be decision readiness.
AI is reducing the cost of knowing
For much of organisational history, information has been expensive.
Reports took time to produce, analysis required specialist expertise and information travelled through layers of management before reaching the people authorised to act on it.
Organisations evolved around those constraints but AI changes the economics.
AI can interrogate thousands of documents, analyse datasets, identify patterns, summarise competing arguments and generate recommendations in seconds.
The obvious assumption is that better and faster information will produce better and faster organisations.
But this raises a problem…
Knowing something faster does not necessarily mean an organisation can act on it faster.
Imagine an AI system identifies a significant opportunity to change pricing but who can approve it?
Perhaps Commercial owns pricing, Finance owns margin, Marketing owns promotions, Technology owns the platform and a senior leadership team ultimately needs to approve anything significant.
AI might produce the insight in thirty seconds but the organisation could still take six weeks to make the decision.
The technological bottleneck has disappeared yet the organisational one hasn’t.
We measure the things that are easiest to see
Most AI-readiness frameworks understandably concentrate on tangible capabilities.
Data quality can be assessed, infrastructure can be audited and skills can be mapped. Governance frameworks can be documented and use cases can be prioritised.
But decision-making is harder - it sits between organisational charts, processes, incentives, behaviours and culture.
- A process diagram may show who is supposed to make a decision without telling us who actually makes it.
- A RACI matrix may assign accountability while five people still believe they have a veto.
- A leadership team may talk enthusiastically about empowerment while routinely escalating relatively minor decisions.
- A governance framework designed to manage risk can gradually become a mechanism through which nobody has to personally accept responsibility for the decision.
These problems already exist in organisations.
AI doesn’t remove them. It exposes them.
What happens when AI challenges the organisation?
There is another dimension to decision readiness that becomes increasingly important as AI capabilities improve.
What happens when the system reaches a conclusion people don’t like?
Suppose an organisation has spent years pursuing a particular customer segment and AI analysis suggests it is structurally unprofitable.
Or perhaps it identifies that a long-established internal process creates little measurable value.
Maybe it concludes that the programme receiving the greatest executive attention is unlikely to deliver the expected return.
The technical question is whether the analysis is reliable… the organisational question is more uncomfortable:
Are we genuinely prepared to change our minds?
Organisations are not rational machines waiting for better information.
Decisions are affected by status, incentives, previous commitments, departmental interests, personal relationships and organisational history.
A sufficiently capable AI system may therefore do more than automate work - it may challenge the stories an organisation tells itself about how it works.
That requires a very different kind of readiness.
Decision readiness has several dimensions
Before asking how AI can enable faster decisions, organisations should understand how decisions happen today: Who actually has authority?
Not who the organisation chart suggests has authority, but who can genuinely make a decision without seeking permission elsewhere.
- Where are the hidden vetoes?
- How many people can stop a decision compared with how many can make one?
- Which decisions are routinely escalated, and why?
Some escalation is necessary; some exists because responsibility is unclear, incentives discourage risk or people have learned that making the wrong decision is more dangerous than making no decision.
How reversible is the decision?
Organisations often apply similar governance to fundamentally different decisions. A difficult-to-reverse strategic commitment deserves scrutiny yet a low-cost experiment that can be reversed next week probably does not.
How quickly can evidence change a decision?
If new information appears, can the organisation adapt or does changing direction require someone to admit that the previous decision was wrong?
Perhaps most importantly:
Who is accountable when AI contributes to a decision?
If a human accepts an AI recommendation and the outcome is poor, who owns the result?
- The user?
- Their manager?
- The technology team?
- The vendor?
- The governance committee?
If that question has no clear answer, people will eventually discover the safest response… don’t decide.
AI could make slow organisations feel even slower
There is a paradox here: the faster AI becomes, the more visible organisational friction may become.
When analysis took three weeks, a two-week approval process did not necessarily look extraordinary, however, when analysis takes three seconds, those same two weeks become difficult to ignore.
This could be one of the more interesting consequences of widespread AI adoption.
We may discover that many of the constraints we attributed to technology were never technological constraints at all - they were organisational ones.
AI may therefore create pressure not merely to automate processes, but to reconsider why those processes exist.
Not simply to give managers better information, but to ask why particular information needed to travel through multiple layers of management in the first place.
Not simply to accelerate decisions, but to question who should be making them.
Before AI transformation comes organisational clarity
None of this means organisations should stop investing in data, governance, infrastructure or AI skills.
All of these remain essential, but none of them answers the organisational question.
An organisation that wants to become genuinely AI-ready should be able to answer some deceptively simple questions:
- What are the decisions that matter most?
- Who has the authority to make them?
- What information do they need?
- What can be decided locally rather than escalated?
- Which decisions are reversible?
- Where do unnecessary approvals and hidden vetoes exist?
- How will AI-generated evidence be challenged?
- Who remains accountable for the outcome?
Those questions don’t require AI to answer but answering them may determine how much value AI eventually creates.
The organisations that benefit most from AI may not simply be those with the best models, the cleanest data or the largest technology budgets.
They may be the organisations capable of turning better information into better judgement… and better judgement into action.
Before asking whether your technology is ready for AI, it may be worth asking whether your organisation is ready to decide.
