If you’ve taken a look at your website stats recently you may have noticed that your traffic has been impacted by the ‘AI-effect’. That’s the impact of people getting the answers using their AI tool of choice rather than coming to your website directly like they used to. While this might seem like something to worry about, declining website traffic doesn’t necessarily mean declining conversion rates, and the gap between those two things is worth understanding.

For years, we've measured websites by how many people showed up. Traffic was the metric, pages were the product, and the goal was to get people in, move them around, and guide them to an outcome.

That model's breaking down. AI answers, comparison tools and agents are swallowing the ‘discovery layer’: the part of the journey where people work out what exists and what their options are.

Your website isn't reliably where people go to explore any more. It's where they go (if they go at all) to decide and act. Websites are becoming less of a destination and more of a ‘decision layer’, the last stop in a journey that mostly happens elsewhere.

This isn't an argument for spending less on your website. It's an argument for spending differently and building something genuinely AI-ready.

The behavioural shift is well underway. Research by Bain found that around 80% of consumers now rely on zero-click results in at least 40% of their searches. It estimates organic traffic will drop 15% to 25% as AI summaries and LLM-based search take over.

Fewer people will land on your site via a generic search and more will arrive already part-way to a decision, having done their early research with an AI assistant, an aggregator or a comparison tool. The ones who do land are closer to action than they used to be.

That raises the bar for what a good website needs to do.

The question used to be ‘how do we get more people browsing more pages?’. Now it needs to become ‘how do we make our organisation easy to find, trust, compare and act on, across all human and machine-led journeys?’.

This points to a different ambition: making your website genuinely trusted, machine-readable and conversion-ready across every channel in your market.

Your website is becoming infrastructure — the trusted layer behind every other channel, including the ones you don't control. It's what AI assistants reference, what comparison tools pull from, and where external services hand people off to when they recommend you.

Be a source of truth. Product details, pricing logic, eligibility rules, support policies, value propositions: it all needs to be structured, accurate, maintained and easy to digest, for human readers and for the systems now reading on their behalf. Messy or outdated content confuses customers, and it confuses the AI models summarising your offer to everyone else.

Make comparison easy. Comparison between you and your competition gets harder when your offer is ambiguous. If your conditions, assumptions and pricing logic aren't clearly structured, neutral tools and agents will either misrepresent you or skip you entirely. Clarity used to be a UX problem, now it’s a discoverability one.

Make handoffs clean. Handoffs from external tools are where more of your conversions are going to happen. When an aggregator or AI assistant recommends you, the path from ‘this one looks good’ to ‘I'm now a customer’ needs to actually work. That means creating good product-specific landing pages, sensible entry points for sign-up or contact, and journeys that hold up under traffic that didn't come through your homepage. If you have a clunky handoff, trust quickly evaporates.

Bridge decision-making and servicing. Your public site and your customer service layer can no longer be separate. People move between them fluidly, and your website needs to do the same, introducing propositions with enough context and reassurance to convert, then passing people into service experiences without friction. The join between the two is where trust is either confirmed or lost.

If you're still measuring success by sessions and pageviews, you're measuring the wrong things.

Better metrics focus on influence and outcome. How often does your organisation get referenced in AI answers? How are you represented on comparison tools? What's the quality of the visits you do get? Are people arriving further down the decision path? What's your completion rate on sign-ups, applications or support tasks? How well do people arriving from those external services actually convert?

Keep measuring your website by traffic alone and you'll conclude it's declining in importance at the exact moment it's becoming more important.

But that importance comes with a catch: most websites aren't yet built to perform in this environment.

Plenty of organisations are doing solid technical work on AI readiness already: structured data, schema, llms.txt files, cleaner CMS outputs, better machine readability. That work matters. It's a good foundation.

But it's not the whole answer. Being visible isn't the same as being usable. External services and future agents also need to be able to use you. That means understanding your offer well enough to compare it fairly, knowing how to route people into the right journey, and handing people off without the experience breaking down. That's a higher bar than "machine readable". It's "machine operable". Getting there means thinking about your content, propositions, comparison channels and journey entry points as a connected system, not a set of separate pages.

The temptation with a shift this big is to try to fix everything at once. The smart move is to find the place where the shift is most visible in your market. A specific aggregator, a comparison tool, or an AI assistant your customers are actually using. Start there and treat it as a focused test or pilot. Audit how you currently show up there. Check whether your content can be extracted and compared accurately. Review the handoff from that service into your own journey. Reset your measurement so you can see what's actually happening.

A focused pilot does two useful things: It gives you a practical response to the shift, in a space where the impact is real and measurable, and it generates the insight that shapes a broader roadmap: one that evolves your website, content model and digital presence for an environment where AI is becoming part of every customer journey.

As discovery fragments across AI answers, comparison engines and agent-led services, your website becomes more important as the place where you explain, reassure, convert and service people well. It's the brand-owned operating layer behind a more distributed journey: the source everything else references, the trusted ground truth in a world of synthesised summaries.

That's a much more valuable role than being a top-of-funnel traffic destination. It's also a harder one, because it demands your content, product data and journeys be more rigorous and better connected than they probably are today. And rigour alone won't do it. People still make decisions emotionally, even when they arrive armed with AI-assisted research. The structured layer has to earn trust; the human layer has to close it.

The organisations that make this shift early will show up well wherever their customers are comparing and deciding.

Websites used to be destinations. Now they're the ground everything else stands on.

This is the work we're doing with clients right now. If you're curious where your digital presence sits in this shift, we'd be glad to take a look with you.