Something has shifted. You can feel it in the numbers, even if you can’t fully explain why just yet. Traffic patterns look different. Content and campaigns that once did well now underperform. The playbook that worked two years ago has begun to deliver diminishing returns.

The problem isn’t your content strategy; it’s that the customer you think you’re targeting doesn’t behave the same way anymore.

Most organisations are still optimising for a version of the customer that’s in decline – someone who searches Google, clicks through to a website, reads a few pages, and gradually moves down a funnel that marketing can see, measure and influence at every stage. That model might have worked for a long time, but it’s no longer how decisions get made.

Instead of using keywords to explore options via search results and website visits, customers are turning to AI platforms like ChatGPT, Claude and Google Gemini, asking detailed questions and jumping straight to the answers.

Even in Google search, everything has changed. With AI Mode having reached one billion monthly users, Google is doubling down on AI search with a massive upgrade to the search box. The new search box expands dynamically as you type, with autocomplete helping to quickly formulate much longer, more nuanced, personal and conversational queries. And by clicking the + icon, you can add images, videos, PDFs and even other active Chrome tabs to provide more context and information to each query.

AI has changed not only where and how customers look for information, but also how they form opinions, build preferences and arrive at shortlists. And most of this activity is completely invisible to your analytics. According to a survey conducted by McKinsey, around 50% of consumers now intentionally seek out AI-powered search when making buying decisions, with 44% describing AI as their primary and preferred source of insight – ahead of traditional search (31%), brand/retailer websites (9%), and review sites (6%).

McKinsey goes on to predict that unprepared brands may see traffic from traditional search channels decline anywhere from 20 to 50% as a result of this shift.

That stat alone should be a five-alarm fire for any business still clinging to the same old, familiar customer journey model.

The funnel made sense – until it didn’t

Let’s be honest: the traditional customer journey was never completely accurate, derived as it was from convenient assumptions that don’t always stack up to reality.

While some journeys may be roughly linear, others can jump all over the place. While discovery often happens via search, it doesn’t always. The researcher, decision-maker and end user might be the same person, or multiple stakeholders with very different agendas, particularly in B2B.

Even so, we needed a framework to hang our strategies on, and the traditional view of the journey or funnel – Awareness → Consideration → Decision – still made sense most of the time.

However, the biggest assumption we need to reevaluate is this: Customers will visit our websites and other assets at identifiable strategic touchpoints along the journey, allowing us to track and influence behaviours, map content and measure performance.

Many organisations continue to think in terms of clicks and touchpoints. But while some customers still behave this way, they’re gradually becoming an endangered species. The traditional model no longer reflects the modern customer, who can explore a need, identify potential solutions, assess your category, develop preferences between competitors, and make a purchasing decision entirely within a single AI conversation – all before a single click shows up in your analytics.

Your brand either features in these AI conversations, or it doesn’t. And if AI isn’t recommending your brand to a potential customer, who is it recommending instead?

Instead of clicks, the new customer journey is driven by brand relevant citations in AI responses. Instead of traffic, the new goal is for your brand and its messaging to be present at every stage of a potential customer’s research.

When it happens (if it happens), that click through to your website is no longer a discovery step. Visits from AI are increasingly about confirming information, about validating and acting on decisions that took shape long before the click occurs.

The invisible decision phase

Think about what this means in practice. Say a small business owner is looking for a new accounting software platform. Instead of Googling a category term and clicking through ten websites to gather details, evaluate features, and compare pricing, they provide AI with a use case and a list of requirements and let it assess all the options for them.

The AI curates information from hundreds of sources, before summarising a handful of credible options as a shortlist. The customer then asks increasingly specific additional questions to whittle that list down further. Each new question and answer also has the potential to throw up fresh lines of enquiry, veering off on tangents that refuse to fit the old, linear model of the customer journey.

It doesn’t matter whether this conversation happens in a single session or in multiple interactions spread over a few days or weeks. Either way, any brand mentioned or recommended in these responses is still completely unaware that this potential customer even exists.

This is the new reality. And the personas and customer journey maps in most organisations simply don’t account for it.

From journey stages to decision questions

In traditional search, the goal has always been to secure the top spot for the most valuable keywords. AI search isn’t about rankings or keywords. Instead, the goal is to effectively own the most valuable questions, to build your content for asking.

1. Review and update your customer personas

Don’t assume these new AI-driven behaviours will be the same for every customer type, demographic, use case or even industry vertical. The highly personalised and detailed nature of AI conversations means you’ll need to re-examine your own customer base with fresh eyes.

Consider the following questions:

  • Which personas are using AI, and how? You may find some customers rely heavily on AI while others are more tentative. Do they use AI as a slightly more sophisticated search engine – typing a simple query, getting an answer and moving on – or are they actively interrogating the responses, asking detailed follow-ups and following links?
  • Which AI platforms are they using? Are they using AI platforms like ChatGPT, Perplexity or Claude, or are they influenced by AI Overviews and AI Mode responses in Google?
  • What problem or need are they trying to resolve? As always, your product isn’t the goal but a means to an end. Understanding that “end”, along with the need or trigger that prompted the search for a solution, is crucial to what comes next.
  • What questions do they ask at each stage? You may have identified potential questions before when mapping customer journeys. However, the nature of AI allows users to ask far more specific and personalised questions, with each new response potentially prompting new tangents.

Unfortunately, there’s currently no accurate way to track and measure genuine AI conversations, and that means you won’t find the information you need in a database or analytics dashboard. That’s why it’s vital your team actively talks to your customers. Only they can tell you how they typically use AI and which questions they’re most likely to ask.

2. Identify key decision questions

As AI conversations don’t necessarily follow a linear path through the same old funnel stages, a more practical approach is to target your content towards key decision questions. These are the questions that not only signal intent but also help the customer to move forward in their thinking and closer to a decision.

Determine which queries might prompt the kinds of responses likely to exert the strongest influence on a customer’s eventual purchase. These are the responses where you’ll want your brand to be cited alongside the most favourable messaging and positioning.

3. Create content designed to answer these questions

Now you can plan your content schedule around answering these various questions, making sure each asset is optimised for LLMs to easily identify and ingest.

However, for an LLM to generate a suitable answer in response to a detailed, highly specific question, it needs similarly detailed and highly specific content to draw from. If your content lacks the necessary depth and detail, the LLM will simply look elsewhere.

As a result, you may find one article won’t be enough to answer each question. You may need to create multiple articles on the same theme, each framed around a different persona or use case and tailoring the advice and information accordingly.

4. Optimise your product schema and data feeds for agentic commerce

For retailers, there’s another aspect to consider – agentic commerce. OpenAI’s Agentic Commerce Protocol (ACP) and Google’s Universal Commerce Protocol (UCP) now make it possible for AI agents to research, shortlist and, in some cases, complete purchases entirely within ChatGPT, Google Gemini or Google AI Mode, without the customer visiting the retailer’s website even once.

Never mind less traffic, agentic commerce means the customer journey can bypass your website entirely.

But unlike brand citations in other AI queries, agentic commerce isn’t interested in your on-page content when deciding whether to surface your product or not. Inclusion is determined entirely by the completeness and quality of your structured data; your product and merchant feeds and product schema.

Our own audit of product pages across 29 major retailers found that 70% were missing all three of the most important schema fields, meaning their products would be ineligible for inclusion. Not ranked lower, not less visible, but excluded entirely from consideration.

This makes structured data and feed optimisation a strategic priority.

The customer has moved on. Will you?

If your content strategy still hangs on outdated assumptions about your customers, you risk publishing content for a journey that is no longer typical, optimising for touchpoints that may no longer exist. And because the usual signals no longer reflect true brand visibility or influence, your success metrics no longer reflect true success.

Yes, revisiting and revising all your customer personas and journey maps might feel like a huge chore when you already have a strategy to execute and targets to hit. But in the long run, executing the wrong strategy could be far more damaging.

Basing your content strategy on an accurate understanding of how your customers actually behave today ensures that everything you publish works even harder.

Need help rethinking your content strategy for the age of AI? Get in touch to arrange a discovery session: [email protected]