If you were building your business today, would it be AI-native?

Most businesses are using AI now. Our research at the end of last year found that 78% of UK businesses were already using it in some capacity, and the conversation has moved quickly since then, with AI tools being adopted across organisations, existing products gaining AI features and teams finding ways to automate work that previously took people hours.

There is clearly value in all of this, but adoption itself is becoming a less interesting measure of progress. The key question is whether businesses are using AI to improve what they already have, or whether they are starting to think more fundamentally about what they would build differently if they were starting today.

‍Improving what already exists

‍A lot of the first wave of AI adoption has understandably focused on making existing things better. Businesses have taken processes that already exist and automated parts of them, given employees copilots, added AI assistants to existing products and found ways to generate or analyse things that previously required significant manual effort.

Some of those changes can create a lot of value and there is no reason every application of AI needs to fundamentally change a business. The limitation comes when every conversation starts with the assumption that the product, service or business we have today is broadly the one we should have tomorrow and the task is simply to work out where AI can be added to make it faster, cheaper or more efficient.

That assumption is increasingly worth challenging. If AI changes what a product can do, what it costs to deliver a service, how expertise can be made available or how customers can interact with a business, then there is no reason the answer should always be a slightly better version of what already exists. In some cases, the more useful starting point is to ask what you would build if the existing version didn’t exist at all.

‍What would you build today?

‍Imagine starting your business today with everything you have learned about your customers and your market, but without the product, processes and technology that have accumulated around it over the years. You know where the friction is, which parts of the service are expensive to deliver, which decisions rely on a small number of experienced people and which parts of the customer experience have always been more complicated than they should be.

The interesting question is what you would build now that the technological constraints have changed. Something that currently requires a service team might become a product customers can use themselves, expertise that sits with a handful of people might become available across an organisation or a customer journey built around navigating a series of steps might disappear because the product can understand what someone needs and do more of the work for them.

Go further and the implications don’t necessarily stop with the product. If AI changes how something is delivered, the cost of delivering it can change too and that can open up different commercial models, different ways of serving customers or entirely new services that previously wouldn’t have made economic sense. At that point, the conversation is no longer really about adopting AI. It is about what the business itself could become.

‍Established businesses have an advantage

‍Startups tend to get a lot of attention in this conversation because they have the freedom to start with a blank sheet of paper, without legacy technology or established ways of working shaping what they build. But established businesses have advantages that a new entrant would spend years trying to create: existing customers, distribution, trust, data, domain expertise and a detailed understanding of how their market actually works.

The challenge is separating those genuine advantages from everything that has simply accumulated around them. A ten-year-old process isn’t an advantage because it has been there for ten years, just as a product architecture designed around technical constraints that no longer exist shouldn’t automatically dictate what gets built next.

For established organisations, the opportunity is to hold onto the things that genuinely make the business difficult to compete with while being much less sentimental about the product, service and operating assumptions that sit around them. That gives them the ability to think more like a new entrant without giving up the advantages that come from already knowing the market.

‍Build the competitor you don’t want to face

‍One way of doing that is to stop asking where AI fits into the existing business and instead imagine the competitor you would least like to enter your market tomorrow. Assume they have access to the same AI capabilities as everyone else but none of your legacy systems, organisational structures or assumptions about how the industry is supposed to work.

Think about what they would build, how their product or service would work, which parts they would automate completely and where people would still add value. Think about what the customer experience would look like if they weren’t required to follow the same journey they do today, how quickly the service could be delivered and what a different cost base might allow that competitor to charge or offer.

Then add back the advantages you already have: the customers, knowledge, data, reputation and distribution that the hypothetical new entrant doesn’t.That creates a very different conversation from finding more places to use AI. It starts with the opportunity to rethink the product, the service and potentially the business around what is now possible, while using the advantages that established organisations already have to make that new version harder to compete with.

The next phase of AI adoption will inevitably involve more tools, more automation and better versions of existing products, but the bigger opportunity is to decide which assumptions are still worth keeping and which decisions are worth making again. For established businesses, the question isn’t simply how much AI they can adopt, but what they would choose to build if they were building their business today.