A client of ours — a small bank — recently showed us something they were excited about. An AI assistant in their online banking platform. When customers log in, the assistant surfaces spending insights: upcoming subscription payments, annual costs, patterns they might not notice.
It is genuinely useful. And it will not be a differentiator for long.
Remember when Uber's real-time tracking felt revolutionary? When seeing your driver on a map, paying automatically through the app — that was remarkable. Now every ride-hailing service does it. It is expected behavior, not a selling point.
The same compression is happening with AI. Gmail now manages your inbox the way third-party tools used to charge for. Banks are rolling out spending assistants. Customer service chatbots are everywhere.
MIT Sloan Management Review argues that AI will not provide sustainable competitive advantage precisely because these capabilities commoditise quickly. If everyone has access to the same technology, it moves the market but does not uniquely advantage anyone.
So what is left when the technology is table stakes?
Brand.
Not the logo. Not the colour palette. Not even the tone of voice document sitting in a shared drive somewhere.
Brand as behavior. Brand as the operating system that determines how every AI-powered touchpoint interacts with a customer.
When a bank's AI assistant tells a customer their Netflix subscription costs $240 a year, what happens next matters more than the insight itself. Does the AI nudge them to cancel? Offer a budgeting tool? Simply report facts and move on?
That response is brand. It has to come from somewhere specific — a set of behavioral specifications that reflect what the organisation actually believes.
In traditional organisations, there are many layers between a strategic idea and its execution. A CEO articulates a vision. It filters through management. Gets reinterpreted by middle managers. Arrives at the front line diluted or distorted.
Every layer of that bureaucracy is a place where inauthenticity can hide. "We wanted to do that, but finance said no." "That didn't make it through the approval process." The gap between intent and execution is wide enough to absorb any amount of vague aspiration.
AI compresses that gap to almost nothing. Whatever you encode is what the customer experiences. There is no middle management filter softening the message. No reinterpretation. If the culture is hollow, the AI will expose it the same way silence exposes a weak delivery process.
Here is where most brand work breaks down.
A CEO saying "we are customer-obsessed" is a value. It sounds good in a town hall. It does not help configure an AI.
"When a customer asks about fees, explain the fee structure, offer alternatives, and never make them feel penalised" — that is a behavioral specification. An AI can execute that. A value, it cannot.
This distinction matters because 80% of companies are expected to adopt AI-powered chatbots and the ones that configure them with values instead of specifications will produce interactions that feel generic. The AI will be polite, but it will not feel like the brand.
This is where smaller organisations have an edge.
A small bank can move faster than a large one. The founder's culture is still alive in the building. The distance between "what we believe" and "how we behave" is short. Encoding that into an AI system is a smaller, more honest project.
A large organisation with professional management — a CEO who inherited the culture rather than built it — faces a harder task. The culture may be documented in brand guidelines, but how deeply it is understood by the people configuring AI systems is a different question.
Smaller organisations can experiment, iterate, and build a database of behavioral specifications organically. By the time incumbents realise what happened, the upstart has already become the customer expectation.
There is another dimension most conversations about AI and brand miss entirely.
If a staff member uses an internal AI tool — to draft emails, manage their calendar, prepare reports — that tool should also embody the brand. The way it phrases suggestions, how flexible it is about scheduling, the tone it uses — all of that shapes how the employee works and communicates.
Gartner predicts that over 20% of workplace apps will use AI-driven personalisation by 2028. A small bank that gets this right has their AI speaking the same language to customers and employees. The teller drafting an email and the chatbot answering a complaint operate from the same behavioral logic. That consistency is difficult to replicate.
The window matters. Right now, most organisations are either experimenting with AI or have not started. The ones that use this period to articulate their brand at the behavioral specification level — not the aspiration level — will have a significant head start.
Because once AI is standard in every customer touchpoint, the only question left is: does your AI feel like your brand, or does it feel like everyone else's?
The technology will not answer that for you. Only the hard work of knowing what you actually stand for will.
This article applies brand-as-operating-system.