There was a fairly familiar rhythm to technology consulting.
The client had a problem. The consultants arrived, conducted discovery, interviewed stakeholders, disappeared for a while and came back with a recommendation, a solution design or perhaps an impressive-looking deck.
Then somebody built it.
That model hasn’t disappeared. But AI is beginning to change it.
Increasingly, clients don’t simply want consultants to tell them what the answer is. They want to explore the answer with them — testing ideas, building prototypes and discovering what’s possible as they go.
And that requires a subtly different set of consulting skills
From AI Experimentation to AI Implementation
Part of the reason is simply that AI adoption is becoming much more mainstream.
The latest ONS figures show that around 35% of UK businesses with 10 or more employees now report using AI, up from roughly 12% in late 2023. In the information and communications sector, the figure is already 58%.
Professional services is moving particularly quickly. The UK Government’s 2026 AI Adoption Plan for Professional and Business Services reports that 43.4% of firms in the sector were using AI by December 2025, compared with 31.4% a year earlier.
So we’re moving beyond the stage where every AI conversation begins with, “What is generative AI?”
The more interesting client conversations now tend to be about what to do with it.
Which processes could we change? Where might AI genuinely improve customer service? What could we automate safely? Where do we still need human judgement? How do we get from an interesting proof of concept to something people will actually use?
These questions don’t always lend themselves to the traditional model of disappearing for six weeks and returning with “the answer”.
Often, the answer needs to be discovered together.
The Rise of Co-Creation
Co-creation isn’t new, of course. Good consultants have always worked collaboratively with their clients.
What’s changing is the speed at which ideas can now be turned into something tangible.
A workshop that might once have ended with a collection of Post-it notes and a promise to return with some recommendations can potentially finish with the outline of a working prototype.
An idea for a new customer interaction can be mocked up. A proposed AI assistant can be tested against real questions. A cumbersome internal process can be redesigned and demonstrated before anyone commits to a major implementation.
That changes the dynamic.
Instead of:
Discover → Design → Present → Build
we increasingly get something closer to:
Discover → Build something small → Test → Learn → Redesign → Build again
This is much more than Agile delivery with AI added to it. The client is participating in the discovery of the solution, rather than simply approving one created elsewhere.
And that changes what clients need from consultants
Knowing the Answer Becomes Less Important Than Asking the Right Questions
Consulting has traditionally placed a premium on expertise — quite reasonably.
Clients pay for people who know things they don’t.
That isn’t going away. But when technology is changing rapidly and the possibilities are not yet fully understood, there may not be a single pre-existing “right answer”.
The consultant’s value increasingly comes from helping the client ask better questions.
What problem are we actually trying to solve?
Which parts of this process genuinely require intelligence, and which are simply inefficient?
What would success look like for the user?
What data would this require?
What happens when the AI gets something wrong?
And perhaps the most useful question of all:
Should we be using AI for this at all?
That last question matters. A consultant who recommends AI for every problem will quickly lose credibility.
Sometimes a workflow change, some basic automation or a better user interface will solve the problem perfectly well.
Good AI consulting isn’t about finding places to put AI. It’s about finding better ways to create value.
The Prototype Is Becoming Part of the Conversation
One of the more interesting consequences of generative AI is that the distance between an idea and a prototype has shortened dramatically.
This matters because people respond differently to something they can see.
Ask ten stakeholders what they want from an AI-enabled service and you’ll probably receive ten slightly different descriptions.
Give them a basic prototype and you’ll hear:
“That’s not quite what I meant.”
Excellent.
You’ve learned something.
Rapid prototyping makes disagreement visible earlier, when it is still relatively cheap to resolve.
It also shifts the role of technical consultants and solution architects. Instead of arriving after the business requirements have supposedly been defined, they can become active participants in discovery.
Technical knowledge becomes part of the creative process.
That is potentially a much more valuable role.
Co-Creation Isn’t the Same as Making It Up As You Go Along
There is an obvious danger here.
“Let’s experiment” can easily become an excuse for weak governance.
A prototype isn’t a production system. Something that works impressively in a workshop may behave very differently when exposed to sensitive data, thousands of users, security controls, regulatory requirements and the glorious complexity of real corporate IT.
This is particularly relevant with AI because the technology can create convincing demonstrations very quickly.
The danger is confusing speed to prototype with speed to value.
Research from McKinsey illustrates the gap. Its global AI research found that workflow redesign was one of the factors most associated with organisations seeing bottom-line impact from generative AI, yet only 21% of respondents whose organisations used generative AI said they had fundamentally redesigned at least some workflows.
The UK Government’s 2026 professional services AI plan identifies a similar issue: AI use is spreading quickly among employees, but many organisations haven’t yet made the corresponding changes to workflows, decision-making and organisational design needed to turn individual productivity gains into business-wide improvements.
In other words, giving everybody an AI tool isn’t transformation.
Changing how the organisation works might be.
The Consultant as Facilitator, Challenger and Translator
All this puts greater emphasis on some very human consulting capabilities.
Facilitation is one.
A good co-creation workshop isn’t a brainstorming free-for-all. Someone needs to keep people focused on the business problem, draw out quieter voices, challenge assumptions and move the group towards decisions.
Another is the ability to manage ambiguity.
Traditional consulting can sometimes create an expectation of certainty: analyse the situation thoroughly enough and the correct answer will emerge.
AI projects don’t always behave like that.
Sometimes the responsible answer is:
“We don’t know yet. But here’s how we can find out quickly and safely.”
That requires confidence — particularly when speaking to senior clients who are paying for expertise.
Then there’s translation.
Technical teams need to explain what AI can realistically do without drowning people in terminology. Business stakeholders need to articulate what value actually means. Risk, legal and security teams need to be involved without becoming the people invited in at the end to say “no”.
The consultant increasingly sits between all these groups.
Not necessarily as the person with every answer, but as the person who helps the organisation reach a better answer.
It Also Changes the Client Relationship
Perhaps the biggest shift is psychological.
Traditional consulting can create a fairly clear division:
We are the experts. You are the client.
Co-creation deliberately blurs that boundary.
The consultant brings technical knowledge, external experience, methods and challenge. The client brings organisational context, domain expertise, customer knowledge and an understanding of what will actually survive contact with their business.
Neither side has the complete picture.
That requires a little humility from consultants.
It also requires more involvement from clients.
Co-creation isn’t code for getting the customer to do the consultant’s job. Successful collaboration requires committed stakeholders, access to the right people and a willingness to make decisions.
If the client says, “We’d like a highly collaborative engagement — we’ll see you again at the steering committee in six weeks,” it probably isn’t co-creation.
What Does This Mean for Technology Consulting Teams?
It means the skills profile of a strong consulting team is broadening.
Technical expertise remains fundamental. Nobody wants an engaging facilitator who doesn’t understand the technology.
But alongside technical depth, increasingly valuable capabilities include workshop facilitation, design thinking, commercial awareness, storytelling, stakeholder management, rapid prototyping and the ability to challenge a client constructively.
There is another important skill too: knowing when to slow down.
AI encourages speed. Good consulting applies judgement.
The best consultants will know when a two-day experiment is preferable to a two-month analysis — and when a two-day experiment absolutely isn’t enough.
That balance between curiosity and discipline may become one of the defining characteristics of successful AI consulting.
Building With Clients, Not Just For Them
AI isn’t making consultants irrelevant.
But it may be making one particular version of consulting less valuable: the consultant whose primary advantage is access to information, frameworks or analysis that the client couldn’t easily produce themselves.
AI is steadily democratising some of those capabilities.
What remains much harder to automate is helping a group of people understand a messy problem, reconcile competing priorities, make sensible trade-offs and turn an emerging technology into something useful.
That is where co-creation becomes interesting.
The future of technology consulting may be less about arriving with a beautifully packaged answer.
It may be about creating the conditions in which the consultant and client can discover the right answer together.
And sometimes that will mean saying:
“We don’t know exactly what the solution looks like yet.
Let’s build enough of it to find out.”