The AI-powered consultancy

Many consultancy leaders see AI as a threat.

In a recent TCGN members’ workshop, Umbar Shakir, former partner at Gartner and Chief AI Officer at Havas, tackled this issue head-on. She said, “there’s a lot of hype out there about consulting being dead, but in reality the value is actually coming into sharper focus”.

This was the context to Umbar’s workshop. It wasn’t about  ‘how do we use AI’, but ‘how do we keep our delivery valuable as the easy parts get automated’. And here are Umbar’s top tips for keeping AI as an advantage, not a risk.

1. The ‘bums on seats’ model will get squeezed first

If your model relies on repeatable production work, AI is going to compress it.

Umbar raised concerns “that the ‘crank the handle bums on seats’ business model is going to be impacted, which will erode your fees. Consultancies that don’t adapt and evolve the type of consulting value they’re bringing simply won’t survive”.

If your model relies on repeatable production work, AI is going to compress it.

The practical takeaway is not to panic. It’s all about prioritisation. Umbar explained that if you want to protect margin, stop treating delivery as a fixed ‘project machine’ and start designing it as a value system. Position yourself as a strategic partner to your clients, with clear human moments and clear machine moments.

2. What clients still pay for is judgement in choppy waters

One of my favourite lines from the workshop is Umbar’s metaphor for where consulting still earns its fee. She said: “we can sail the boat in choppy waters, and that’s where the value comes”.

AI is excellent in calm waters. It struggles when the situation is messy

AI is excellent in calm waters. Summarising, formatting, extracting, drafting. It struggles when the situation is messy, political, ambiguous, or emotionally loaded. The bits that makes a senior consultant worth paying for.

So, if your delivery currently looks like a conveyor belt of outputs, this is your prompt to redesign it. Make the ‘choppy water’ moments more explicit:

• the point of view you bring (not just the analysis you ran)
• the decisions you help leaders make (not just the deck you produced)
• the risks you help them avoid (not just the plan you documented).

3. The consultant archetype is shifting, whether we like it or not

A big theme in Umbar’s session was that roles have to evolve, not disappear.

She pushes back on the lazy idea that “AI means we do not need analysts anymore”, instead she asks: “How did we become senior strategists and advisors if we hadn’t cut our teeth on the analyst roles?”.

Then she explains how the ‘senior’ archetype changes, too. In Umbar’s view:

• from an advisor perspective, you need to be more of a creative sparring partner who brings ideas clients don’t already have
• for strategists, it’s  about being a pathfinder, getting into that critical path of delivery
• for delivery management, it’s about being the troubleshooter and the fixer.

That’s a strong prompt for anyone leading a consultancy team. If your career paths still reward output volume over judgement, you are going to feel the squeeze harder.

4. Clients can smell ‘lowest common denominator’ work

This came through strongly in the discussion with members on the workshop.

One consultancy leader shared that some clients are not ‘AI ready’ and, more importantly, they are suspicious of firms over leveraging AI, seeing it as lowest common denominator… output.

Umbar agreed, and her response is a useful line for consultancy leaders to use to build around during client conversations: “we use AI to help us be even more human, where it matters, not generically less human”.

She also gave a great example of what ‘human in the loop’ actually looks like. She talked about using AI for analytics, but still doing the deep work herself: “I make sure I read every single document myself.”

Why? Because clients test you in the room. And as she put it, “you simply can’t rely on AI for everything, clients see right through that”.

That is the real risk for consultancies right now. Not that AI will replace you, but that lazy use of AI will cheapen you.

5. A better delivery play is ‘assistive’ AI, not ‘autonomous’ AI

AI becomes a delivery accelerator, rather than a delivery replacement.

There’s a lot of hype around ‘agentic’ and ‘autonomous’ AI. Umbar’s advice is to keep your head.

She says, “only 1% of organisations have implemented autonomous AI solutions”.

Her stance on delivery is clear: “very, very little of what we do should go fully autonomous. We’re in the realm of it being advisory and assistive for us right now”.

That framing is helpful because it turns AI into a delivery accelerator, rather than a delivery replacement.

6. The practical opportunity: compress discovery to recommendations

What would take a couple of days is now down to a few hours.

This is the bit consultancy leaders should pay attention to because it hits utilisation, delivery speed and consistency.

Umbar explained how her teams use AI across discovery and analytics: turning unstructured client inputs (interviews, documents, strategies) into themes, then refining in their tone of voice, then producing evidence-backed recommendations.

Her headline on impact is the one that really resonated with the members: “what would take a couple of days is now down to a few hours”.

But crucially, she’s explicit about what it’s for: “this isn’t about getting rid of headcount. It’s about being able to service more clients with the same level of staff”.

What to do next

If you want a ‘week one’ version of this, keep it focused on delivery value, not tooling.

• Identify one delivery phase where you lose time (discovery is usually the first suspect).
• Design a simple ‘assistive AI’ workflow that speeds that phase up, without skipping human judgement.
• Define what ‘human in the loop’ means in your firm (what must be reviewed, validated, rewritten, stress-tested).
• Update your delivery story so clients understand where AI is used, and why it improves quality rather than reducing it. And importantly, where it isn’t used.
• Review role expectations for your team, then train towards the archetypes Umbar described, creative sparring partner, pathfinder, troubleshooter.

That’s how AI becomes a delivery advantage, not a credibility risk.