How brands can stay visible as AI disrupts the customer journey

In a recent article for Campaign UK, Iva Johan, Chief Strategy Officer at Bernadette, VCCP’s Digital Experience agency, delves deep into the topic of how AI is marking a collapse of the funnel attribution and how people decide whether to buy items today.

Nine months ago I stopped running my life and let AI take over. As an experiment, I outsourced nearly every decision to AI. What to cook, how to word an awkward email, where to invest money, what to pack for a trip, how to save a dying plant (spoiler, it told me to kill it). I turned ChatGPT into my personal assistant and because she was such a consistent presence in my life, I lovingly named her G.

At first, I wanted to run this experiment to find out how AI might change customer behaviour, but it quickly became something more. I deliberately consolidated around just two tools: one general assistant and one search replacement, not dozens of apps, not every shiny branded agent. I did this to test my hypothesis that you can do a lot with just one assistant. Most people wouldn’t juggle 50 AI agents in their pockets, in the same way they wouldn’t use more than one internet browser: they pick one and stick with it. Branded agents will exist where it makes sense, like banking or telcos, but a holiday planning agent? That’s simply not needed when one assistant will be able to plan holidays, book dinners and even remind me to take the rubbish out before I leave.

What I discovered wasn’t just that AI could save me time or write better emails for me: it completely changed how I choose and decide. The whole process from intent to decision was compressed into a single conversation: no more browsing websites or scrolling through reviews. That shift has huge implications for how brands get discovered, considered and chosen.

The collapse of the funnel

For years we described customer journeys as a funnel: awareness, consideration, decision. With AI, that funnel is collapsing. My AI-powered journey went from "help me plan a week in Sicily" to a complete itinerary with a packing list and dinner recommendations, increasingly with booking of said dinners too. All in one conversation without leaving the chat.

The steps haven’t disappeared per se, they’ve just changed. My assistant was doing the awareness, consideration and decision work I used to do myself, only faster, and invisibly. Even the way I asked questions changed: instead of: “What are the best trail running shoes for women?” I’d ask: “What are the best trail running shoes for me? I run 40-70km a week on non-technical trails and have ankle issues.” That’s a different type of search; contextual, specific and increasingly personal. A lot of my running and training habits ChatGPT knew because it also gave me my training plan, and helped me "diagnose" my ankle issues (for which I did go to a specialist).

With my changing queries, the sources used to generate the responses changed. Rather than glossy brand websites, AI pulled from community discussions and user experiences. As it learned my preferences, even generic questions like “what should I wear to a black tie event” came back tailored to my style. That's a profound shift: conversational AI is training customers to search differently, and those new patterns don’t favour traditional brand touch points.

The wake-up call

Living with AI as my daily decision-making partner isn’t some distant future, it’s here and it’s addictive. Once you experience compressed decision-making, the old way feels slow and clunky. And I was only using today’s conversational tools with limited autonomy and transactional capability. Imagine what happens when agents can truly act on our behalf.

Where service interaction is essential, we will need dedicated branded agents, but in most areas a future in which people rely on a single generalist assistant is imaginable. How much control they relinquish to that agent is arguable, not everyone will be as comfortable making investment decisions based on ChatGPT’s recommendations as I did for the sake of the experiment, but some consolidation of the tools we will use is likely. The companies that figure out how to remain visible now in increasingly AI-mediated journeys before the shift goes more mainstream will have a real advantage.

Three challenges for brands and how to solve them

Your brand has more ways to be found

Traditional search engine optimisation is no longer enough. We now need to optimise for AI systems, from ChatGPT to Google’s AI Overviews to feature answer boxes that surface information in fundamentally different ways, and decide what people see before they ever click a link. Optimising for AI is more complex because content must work for two audiences: humans who want emotion and stories, and AI systems that need structured, comparable data. If AI can’t parse your differentiators, you won’t appear in conversations that matter.

Brands need to proactively audit how AI agents and assistants currently interpret their brand, content and products. This involves systematically testing common queries to benchmark your current representation. Based on these findings, a targeted content strategy can be developed using structured data and clear language to correct inaccuracies and amplify key USPs, ensuring your brand's unique value is clearly communicated to both human and machine audiences.

Experience is shifting from visual to behavioural

With fewer visual movements in conversational interfaces, no screens, no buttons, no pages, brands are defined by their tone of action. Are you fast or slow, proactive or reactive? If your AI agent makes a booking on a customer’s behalf, your brand is speaking through what it does, not just what it says.

This requires brands to define and codify their unique and ownable behaviours – their "tone of action". Go beyond visual style guides and establish principles for how your digital services behave and act: are they swift and efficient, or deliberate and reassuring? Identify critical moments in the user journey – from initiating a request to its successful completion – and intentionally design these interactions to be signature brand experiences. These behavioural touch points, not just visual elements, will become your most memorable differentiators.

Every asset must work for both people and machines

This is the new design challenge. Websites, content and resources must be presented in ways that work not only for human experience but also as inputs for AI systems that retrieve, interpret and act on them. This dual-design challenge means rethinking how you present everything from product information to company values. At its extreme, websites might become “APIs with a skin”, human facing layers on top of machine readable foundations.

The practical approach is to decouple your core information from its creative expression. This means creating a base layer of clear, structured and unembellished content designed specifically for machine interpretation – prioritising raw data and factual accuracy over brand language and tone of voice. In this new landscape, the integrity and validity of your information is essential. Your brand's integrity will be built on the quality of these machine-readable foundations, which, in turn, underpin the inspirational content that your human audience sees.