AI Is Not the Strategy. Better Thinking Is.

How we are using AI at PROBOS to improve thinking, judgement and execution.

AI is having a difficult moment. Depending on what you read, it is going to transform productivity, destroy jobs, undermine expertise or make many of the decisions humans currently make.

Some of those concerns are legitimate. But between the evangelists and the doom merchants, something much more practical is happening.

At PROBOS, we are using AI to help us think better, remember more, challenge assumptions and turn complex information into action more quickly.

That is a very different proposition from simply automating work.

AI can make bad thinking faster

Technology tends to amplify the system into which it is introduced. Put AI into a poorly understood process and you may simply produce confusion more efficiently. Give it weak information and it can produce beautifully articulated nonsense. Ask the wrong question and it can give you a sophisticated answer to the wrong problem.

So our starting point is rarely: ‘What can we automate?’, it is: ‘Where could AI help us make a better decision or execute more effectively?’ That distinction matters.

From assistant to thinking partner

Something interesting happens when an AI system works with you for long enough. It begins to understand how you think, especially if you give it enough information.

In our case, it has meant teaching it how PROBOS approaches a problem. We give it frameworks, operating principles, evidence, previous decisions and rules about how we want information assessed. It learns that we care about commercial outcomes , that activity is not the same as progress, that evidence should be separated from assumption and that an elegant strategy without a route to execution is largely useless.

And, importantly, that it should challenge our thinking rather than simply agree with it. Over time, this creates something much more useful than a generic AI assistant. It becomes a thinking system shaped around the way we work.

Taking that thinking into client work

We are now using this approach within complex client assignments. The examples here deliberately combine and generalise several situations so that no individual organisation can be identified.

A typical assignment might involve strategy documents, presentations, meeting transcripts, interviews, existing processes, competing opinions and hundreds of individual pieces of evidence.

AI helps us interrogate all of that to assess:

• What do we actually know?

• What are we assuming?

• Where do different stakeholders disagree?

• What has changed since the last conversation?

• What question remains unanswered?

• Does the proposed solution actually address the problem we identified?

Instead of spending valuable client time rediscovering information, we can concentrate on the decisions that matter.

When AI starts to work while we are not looking

The next step is particularly interesting. Once the thinking process is clear, parts of it can be automated. We have built AI driven workflows that monitor new information, compare it with what was previously understood and identify what has changed. Before a meeting, an automated process can review the evidence and identify the questions that now matter. Afterwards, another can compare what was said with the previous position, identify contradictions, capture new learning and highlight decisions or actions.

The important point is not that the automation is clever. It is that the automation has been designed to apply the same discipline that we would apply ourselves. It knows what to look for because we have told it how we think. That does not remove the consultant, it gives the consultant leverage.

AI does not get the final vote

There is an important boundary. AI can be wrong. Worse, it can be wrong while sounding extremely convincing. Humans, of course, have their own weaknesses. We forget things, favour our own hypotheses, overlook contradictions and occasionally hear what we expected somebody to say rather than what they actually said. AI has a different set of weaknesses. That is precisely why the combination can be powerful.

We challenge the AI. The AI challenges our reasoning. The client challenges both.

Human judgement remains accountable.

What clients are really buying

Clients do not need PROBOS simply to give them access to AI, it's not what we supply and they can buy that themselves. The value is in what surrounds it:

• Better questions.

• Better evidence.

• Faster understanding.

• Continuity between conversations.

• More rigorous challenge.

• Organisational memory.

• A clearer path from information to decision to action.

Ultimately, that is the point. AI has no commercial value simply because an organisation uses it.

Its value appears when it helps people make better decisions and execute them more effectively.

The opportunity is bigger than automation

Perhaps the most interesting question is therefore not: ‘How much work can AI replace?’

It is: ‘How much better could our people think and perform if AI genuinely understood how our organisation works?’ Its objectives, commercial principles, accumulated knowledge, decision criteria, ways of working and the way its best people think. That is where we believe things become genuinely interesting. Not artificial intelligence replacing human intelligence but human intelligence, organisational knowledge and AI working together.

Bring us a live problem

If you are curious about AI but do not want another generic conversation about ‘AI transformation’, bring us a real problem. Something complicated, something where there is too much information, competing opinions, unclear evidence or a decision that has become difficult to make and Probos will start with the problem, not the technology. The objective is not to use more AI but to think better, decide faster and execute more effectively.

At PROBOS we believe leadership is not becoming simpler. But it can become clearer.

We help leaders navigate complexity with clarity and confidence.