Small AI-empowered teams will be highly disruptive in 2026

AI scales the impact of talented people exponentially

Executive summary:

Every consultancy faces the same impossible trade-off: grow and dilute quality, or stay small and leave opportunities on the table. I’ve lived this tension for 18 years. What I didn’t see coming: AI doesn’t make this trade-off easier to manage. Instead it eliminates the trade-off. Small teams now have exponential advantages.

The consulting problem: talent doesn’t scale.

I’ve been leading creative companies for over 18 years. First as a managing director of the ad agency Boondoggle Amsterdam, later as a co-founder of SUE, SATA and Crowdrock. The business model of every consultancy or agency depends on one thing and one thing only: You can only be successful if you develop a market reputation for delivering distinctive, high-quality work in your domain. That might be a marketing strategy that drives predictable revenue growth, a public support strategy that builds or defends a company’s reputation, or a change strategy that aligns people around a shared mission.

One of the hardest scaling challenges is maintaining consistent quality as the company grows. The faster you scale, the more you’re forced to hire inexperienced people to get the work done. That inevitably puts pressure on your senior staff. They’re accountable for quality across more projects, while also mentoring juniors into real craftsmanship. Not every senior is a natural mentor. Not every junior is a fast learner.

Talent and experience don’t scale. One excellent senior will outperform ten juniors on complex problems, because they know how to move from ambiguity to insight, from insight to strategy, and from strategy to stakeholder buy-in.

You can’t “outspend” this. Seniors have finite time.

Historically, you had two options: hire more seniors (risky, because fit isn’t guaranteed) or invest in training programs (expensive and slow, often taking years before clients trust juniors with high-stakes work).

Then AI arrived, and suddenly, one senior could do the work of ten. The economics of consulting just flipped.

From Mediocristan to Extremistan

There’s a lot of debate currently on whether AI delivers on its promise for transforming the workplace. A study quoted frequently last year was MIT’s Project Nanda. The study found that 95% of all pilots delivered no measurable P&L impact. That’s a spectacularly underwhelming number.

My take: This number is completely irrelevant.

In his book “The Black Swan,” Nassim Taleb introduces the concept of Mediocristan versus Extremistan. Mediocristan is a metaphor for domains where adding one extreme outlier to a population wouldn’t dramatically change the average. If you added the heaviest person who ever lived to a random group of 1,000 people, the group’s average weight wouldn’t be affected that much.

However, if you added Elon Musk to a group of 1,000 people and measured their average income, it would dramatically transform the number. That’s Extremistan.

AI is taking consultancies from Mediocristan to Extremistan.

The MIT study assumes AI impact distributes evenly. That’s classic Mediocristan thinking. But in Extremistan, the 5% that works doesn’t just succeed marginally better. It creates exponential advantages that make old models obsolete.

AI Gives Talented, Entrepreneurial teams Superpowers

One of Steve Jobs’s most famous quotes was “Real Artists Ship.” He was obsessed with turning technology into desirable products and figuring out ways to build scalable production capacity for these products.

There are now multiple labs - Safe Superintelligence, Thinking Machines Lab, OG Labs, and many more - with sometimes no more than twenty people that got more than a billion in funding and $3-35 billion in valuation. The reason is simple: A small AI-empowered team can break through all the barriers for turning opportunities and ideas into scalable products. You now only need a small dedicated team.

Senior people carry mental models: ways of spotting patterns, testing hypotheses, stress-testing ideas, and making trade-offs under constraints. They know how to dissect a messy problem into insights, opportunities, hypotheses, strategy, concepts, tactics. Plus they know how to persuade, reassure, and align stakeholders around a winning direction. They also have a refined sense of what differentiates quality from mediocrity.

LLMs amplify all these capabilities, fast.

In practice, I can:

• Generate synthetic personas based on client data and use them to test and co-create ideas

• Build a project context that “loads” the AI with the relevant constraints, history, stakeholders, and previous decisions

• Prototype strategy quickly using image/video models, thereby reducing weeks of iteration to hours

• Create custom GPTs for training (e.g., leadership simulations) that push back, challenge, and coach in real time

• Let Claude Code build real-time dashboards that track the company’s health and signal potential problems on the spot

There are many ways LLMs give seniors superpowers across the workflow. But the biggest shift is this: it massively increases capacity to execute and adapt strategy.

We used to sell research and advice. A lot of that will commoditize. What won’t commoditize is making the work land. Creating wins, not just documents.

Comedian Jimmy Carr had a sharp line when someone asked if AI can replace him: “No. My job is not jokes, it’s joy.”

That’s a useful analogy for consultancies and agencies. The job is not reports. It’s outcomes.

It’s the 1999 moment, all over again

There’s a LinkedIn post by Craig Taggart that was shared with me multiple times during the last couple of weeks. It’s about how McKinsey lost 3 Fortune 500 clients to small 12-person boutiques¹.

The reason: All the stuff that McKinsey used to charge millions for, only to get an army of junior consultants to do the groundwork, can now be done by a couple of AI-savvy consultants using smart agents for analyzing all the data that matters.

The real value shifts from being a consultant towards being a trusted advisor that helps you prototype, test, execute, and adapt strategy until you achieve your objective. In this new Extremistan world, extreme AI-savviness combined with extreme advisory skills are the new gold.

As an entrepreneur, I feel it’s so much fun to be part of this new era. I’m old enough to remember the excitement in 1999 when the internet had its breakthrough moment. There was this 5-8 year period where the incumbents where trying to figure out how to compete in this new world, but it were the innovators and early adopters who created this new world. This is 1999 all over again.

Related reading: If you want to dive deeper into how organisations can successfully navigate AI adoption, check out this article on AI adoption in organisations. And if you’re curious about how behavioural science can amplify AI’s potential, read more about combining AI and behavioural insights.

Tom De Bruyne

I am a Belgian, living in Amsterdam. I’m founding partner at SUE Behavioural Design Academy and SUE & The Alchemists boardroom consultancy. If you want to follow me: My main publication channel is Linkedin. I published the book “Gamechangers: How small interventions, unlock massive change in 2025 (in Dutch). You can hear me in podcasts (currently all in Dutch) Gamechangers, Wat Gebeurt er Eigenlijk in je hoofd, and Zo Gaat Dat Dus (upcoming). You can also just drop me an e-mail.

1. The post went viral, but the claims were anecdotal, so I cannot verify the validity of the claims he made.