Like most people, I’ve been experimenting with AI for a couple of years now, primarily with Perplexity and ChatGPT. And that is, more or less, what most established non-tech businesses have been doing as well when I talk to them - trying to facilitate AI adoption and adaptation.

And like many people, I mistakenly saw AI as an exciting new tech tool to apply to existing business processes that would drive efficiencies, productivity and insights.

That was an error on my part.

AI will do all that, but it may well also challenge and disrupt the very essence and structures of how we work, our current business models and assumptions and even markets that have been around for a hundred years. It will upset almost everything, everywhere.

Just because a current organisation, an industry, or a sector currently ‘operates’ in a particular way, because ‘that is how it’s done’, will not protect them from impact. Think of the arrival of the motorised vehicle, but faster and greater.

The AI Gordian knot.

There is a risk for organisations that resist AI and for those that embrace it.

Take the partnership model - consultancies, accountancies and law, etc - the career trade you make when you join is that you commit your loyalty, fealty, time, and your best efforts. If you bring enough value over time, you move to the top (surviving the ‘out’) and are rewarded accordingly. If you were smart and hardworking, it was a model you could trust your future on, and everyone could understand the trade.

AI will upend this trade and the very trust that keeps it together, as AI doesn’t work on a time-for-money exchange model; it operates on an output basis and how you can use and leverage AI.

How long a person has been in the business, and even what you know, which used to drive perceived value, will soon become irrelevant because all information and knowledge are now commoditised.

What used to be an advantage may now turn out to be a disadvantage. What used to be a moat becomes a watery prison or barrier to change.

But for many established businesses, markets and organisational structures, the way they reward and recognise will stop them from being able to adapt and change quickly enough to survive. Most organisations are designed and structured for a very different reality.

The question is, can they change?

How AI will impact established businesses and how to respond

Most large-scale legacy or established businesses are built and designed on assumptions that worked in the old pre-AI world order, and as they try to adapt, they be attempting to apply old world order thinking to a new world order event.

For instance, if I were a partner and I had grafted, navigated and compromised my life over 20-30 years to make partner. Only to be told that all those ways of working, with all that blood, sweat, tears and compromises, were not going to be valued in the same way. That I might have to work radically differently, and be rewarded differently as a result, I think I would resent and resist AI - wouldn’t you?

But in that process of resistance is a hidden existential threat to the business, and it's on the clock, as markets change and evolve.

In short, I see an existential threat to most established businesses and operating systems that were designed on a set of assumptions and principles which may no longer stand up.

But it’s not just partnerships, it will challenge most traditional organisations. AI will reward, but also punish different organisations, services, behaviours and capabilities in different ways.

AI’s impact on large-scale, established businesses is likely to be transformative and disruptive on multiple fronts — from their cost base and operating model to how they create, capture, and defend value. Here are some of the ways that AI might impact work and how organisations and their leaders might respond:

Market shifts

• Commoditisation of information and analysis – Businesses that sell expertise, insights, or analysis risk margin erosion as AI makes these faster, cheaper, and in many cases, good enough for customers.
• Shift from service to solution – Clients will expect outcomes, not hours. AI can automate delivery, forcing incumbents to rethink pricing away from time, toward performance-based approaches.
• Loyalty moats dry up – AI will make it easier for customers to try competitors, accelerating churn if loyalty is based on convenience rather than differentiated value.

Operating Models

AI will potentially rewire the core operating model of established organisations.

• Collapse of middle layers – Many traditional management roles (coordination, reporting, approval) will be automated, flattening organisational structures.
• Small heavily utilised HQ functions – Finance, HR, legal, and procurement will be radically smaller and AI-augmented, shifting resources closer to revenue generation. Moving backwards and forwards and being employed in a hyper-targeted way. Think spider web operational system.
• Hyper-personalisation at scale – Products, marketing, and customer interactions can be dynamically tailored, challenging standardised offerings. The off-the-shelf approach is over. Which makes cost management and scale harder
• Faster iteration cycles – Product and service updates will move from quarterly to continuous, requiring new governance and decision-making models

Talent definitions

• Reduction in “process” roles – Repetitive, rules-based work will be automated, reducing headcount in admin-heavy areas.
• Rise of “AI orchestration” roles – New skillsets will emerge in prompt engineering, model fine-tuning, and AI workflow design.
• Redefinition of leadership – Leaders will need to focus less on task oversight and more on strategic clarity, culture, and change management.

Culture

• Resistance to role change – Established businesses will face pushback from employees whose roles are being redefined or eliminated in ways that are currently hard to understand.
• Trust & adoption gap – If leaders don’t align AI adoption with a clear performance agenda, AI will become another underused “transformation project,” with the AI talent leaving in frustration, making it even harder for traditional organisations to compete.
• Loss of tacit knowledge – Over-reliance on AI without human oversight risks eroding critical domain expertise over time.

Competitive landscape

• Moat erosion – AI tools will give smaller, newer entrants the same capability as established players without the legacy cost base. The very benefits of being large and established become the problem, and when startups are valued at higher rates, pay better and have greater profits, what is the point of the old people-heavy model?
• New categories – Entirely new products and business models will emerge, challenging incumbents’ market share before they can adapt.
• Speed and agility – Large organisations’ scale could become a liability if they can’t adopt AI at pace due to bureaucracy or risk aversion.

Solutions

Leaders need to ensure they are doing the following:

• Re-align the commercial model – Leaders will be forced to decide what remains high-value human work vs. what becomes AI-enabled and reprice accordingly.
• Organisational alignment – Ensure AI adoption serves the growth strategy, not just individual department agendas, which will create more friction and drag and stop the performance flywheel
• Accelerating AI Upskill – Build AI literacy across all leadership levels so decisions are well-informed.
• Trust and differentiation – Lean into brand (brand recognition has real value still), relationships, and cultural capital, especially trust, are areas AI cannot easily replicate.
• Hard Decisions – Address redundant layers and roles before market pressures and competition force your hand.

AI is creating a double bind, a Gordian knot for businesses and their leaders. If they try to resist the tsunami of change, they may be washed away, but if they embrace it, they may find it eats their model and market from the inside out …

Yesterday, a team of 12 took 3 months to deliver a project; tomorrow, the same project can be achieved by 3 people in 12 hours through leveraging AI.  AI doesn’t work like a shovel to improve productivity; it allows for hyper-leverage.

Current data suggests that ‘AI transitional displacement’ is forecast to be 7%. But I just cannot see it being such low numbers.

Tinkering won’t cut it

In response, leaders are trying to buy themselves some time by applying old-thinking approaches to cost management, including:

• Applying AI to the more obvious automated roles or repetitive tasks
• Not taking on younger people in particular via graduate programmes and the like
• General slowing down in recruitment, accelerated by not replacing leavers
• Applying AI to Back office roles and functions

But this is merely tinkering. Kicking the can down the road…

At some point soon, most large-scale office-based businesses will have to become AI businesses, and all leaders will need to be AI-EXCOs, or AI-CPOs, and all employees will have to be AI-enabled employees, if they want to keep their jobs or demand higher salaries.

Like it or not, change is coming. Embracing it may be the only route to survival.