4 min to read
Something strange happened in a client review meeting three weeks ago.
A B2B services company — one of our clients here at Codedesign — showed me their analytics dashboard. Organic traffic was down 19% year-over-year. Conversion rate from organic had jumped 41%. Qualified leads were up. The math didn't make sense at first glance.
But when we dug deeper, the pattern became obvious: the people reaching their site were already decided. They hadn't just done research. They'd had research done for them — by AI.
Welcome to the invisible first stage of the buyer journey. And it's happening to your business right now, whether you know it or not.
The Research That Happens Before Your Website Is Ever Visited
Here's what I mean by "invisible."
When a marketing director at a mid-size company starts researching digital marketing agencies, the journey used to look like this: Google search → scroll through results → visit 5–6 agency websites → shortlist 2–3 → request quotes.
In 2026, it increasingly looks like this: Ask ChatGPT or Perplexity "what's the best digital marketing agency for B2B companies in Europe" → get a pre-qualified shortlist → maybe visit one or two sites to confirm → request a quote.
AI-powered search for B2B research queries has grown over 300% in 18 months. Google's AI Overviews now appear in 58% of all searches, with AI Mode sessions ending without any external site visit 75% of the time. These aren't hypothetical future scenarios. They're happening right now, in Q2 2026.
The problem isn't that AI is doing the research. The problem is that most marketing strategies weren't built for a world where an AI system makes the shortlist — not a human scrolling through a results page.
What AI Systems Actually Look For When They Evaluate Your Brand
This is where it gets really interesting — and where most marketing directors are getting it wrong.
AI systems like ChatGPT, Claude, and Perplexity don't evaluate brands the way humans do. They're not impressed by beautiful websites. They don't respond to clever taglines. What they look for is evidence of expertise — and they find it in very specific places.
First: authoritative content depth. An 800-word blog post about "digital marketing trends" won't get you cited. A 3,000-word analysis of why B2B companies in regulated industries are losing organic traffic — complete with data, examples, and a clear point of view — has a real chance of being referenced.
Second: brand signal consistency across the web. When an AI system evaluates your brand, it's reading your website, your LinkedIn page, your reviews on G2 or Clutch, and mentions in industry publications — all at once. Inconsistency is a trust signal failure. If your website says one thing and your LinkedIn says another, AI systems don't know which version to trust. So they move on.
Third: expert attribution. AI systems heavily weight content that includes attributed expert opinions, original research, or proprietary data. This is why at Codedesign, we've been moving clients toward content that features real client outcomes, named expert perspectives, and specific case study numbers. It's not just good writing — it's AI-readability.
I've been following this theme on Voice of Experts, where recent conversations make it clear that the gap between AI capability and actual business adoption is still enormous. As the latest piece there puts it: most businesses are still living in 2023. Which is, frankly, the opportunity.
What We Learned From a Client Who Got This Right (By Accident)
Back to that B2B services client.
When we analyzed why their qualified lead quality had improved despite lower traffic, we found something telling: they had, somewhat accidentally, built exactly the content profile that AI systems favour.
Over the previous 18 months, they had published 12 long-form guides (all 2,000+ words), contributed expert commentary to three industry publications, and collected 34 detailed client case studies with specific outcome metrics. They weren't trying to optimise for AI. They were just trying to be genuinely useful.
The result? When we tested their brand name and service category in ChatGPT, Perplexity, and Google's AI Mode, they appeared in the top three recommendations 68% of the time. Their competitors — who had invested heavily in Google Ads and social media — appeared 11% of the time.
The leads coming in through AI-influenced channels were arriving with a pre-formed sense of trust. They'd already been told — by an AI they trusted — that this company was credible. The average sales cycle shortened by 3.2 weeks.
This is the compounding advantage of building for AI visibility: once you're in the reference set, you tend to stay there — because AI systems are conservative about changing their recommendations without a strong reason to. The flip side is equally true: if you're not in that reference set today, getting in becomes harder every month that passes.
For businesses that want to understand where they stand, our team at Codedesign has been running AI visibility audits as part of our strategy engagements — mapping how brands appear across the major AI platforms and identifying the content and signal gaps that need to close.
Three Practical Moves Marketing Directors Can Make This Quarter
I don't want to leave you with just the analysis. Here's what actually matters right now.
Audit your AI footprint. Search your brand name, your category, and your key services in ChatGPT, Perplexity, and Google AI Mode. Note whether you appear, how you're described, and whether the description is accurate. This is your baseline. Do it today and update it monthly.
Invest in depth, not volume. If you're publishing eight 600-word posts a month, consider publishing two 2,500-word pieces instead. The calculus has changed. AI systems reward genuine expertise signals over publishing frequency. One definitive guide beats ten shallow takes, every time.
Build external citation. AI systems cite sources that other credible sources also cite. This means getting your executives featured in industry publications, contributing to research reports, earning backlinks from authoritative domains, and managing your review presence on G2, Trustpilot, and Clutch. These signals aren't just for traditional SEO — they're now direct inputs into how AI systems evaluate your credibility.
I've been thinking about this shift through the lens of Ethan Mollick's Co-Intelligence, which argues that AI changes the underlying nature of every professional workflow — not just the tools inside it. For marketing, the implication is direct: when AI is part of your prospect's research process, it becomes part of your marketing funnel. Whether you invited it there or not.
The question isn't whether this is happening. It is. The question is whether you're going to shape how AI describes your brand, or leave that to chance.
What's your experience? Are you seeing shifts in where your best leads are coming from, or how informed prospects are when they first reach out? I'd genuinely love to hear what patterns other marketing directors are noticing — share your thoughts in the comments, or reach out directly through Codedesign.
Add comment ×