How I think · AI search
AI search didn’t change what good content is. It changed who reads it first.
For most of my career, the first reader of a B2B page was a person typing a question into Google. Now it’s often a machine.
By Meryl D’Sa-Wilson
Before a buyer ever lands on our site, ChatGPT, Perplexity or Google’s AI Overviews has read the page, decided whether it’s worth citing, and summarized it, sometimes well and sometimes not.
That sounds like a reason to start over. In my experience, it’s mostly a reason to do the fundamentals better.
AI tools cite the brands they already trust
When we started tracking AI referrals, the sites that kept showing up weren’t the ones with the cleverest pages. They were the ones that had covered a topic thoroughly for years. Topical authority, clear entity signals and depth mattered more than any single page.
We learned this the hard way. One competitor kept outranking us in ChatGPT answers. Their edge looked like stronger third-party links and mentions, not better pages. So we updated our pages to match theirs in depth and structure, and then shifted our efforts towards improving the overall "trustworthiness" factor of our website.
Each channel reads a different part of the funnel
The biggest surprise was where the citations landed. ChatGPT and Perplexity were sending buyers straight to our pricing page, about as bottom-of-funnel as it gets. Google’s AI Overviews cited our educational guides instead.
So we stopped treating "AI search" as one channel. Pricing and signup pages got clearer plans, plain-language comparisons and fewer steps. Educational content got direct answers near the top, followed by the depth a human reader still needs.
What I’d tell any B2B content team
- Answer the real question early. Start with the answer buyers ask for, in their words, then explain. I’d rather overexplain than leave a gap an AI tool fills with a competitor’s page.
- Structure for skimmers, human and machine. Clear headings, tables, FAQs and structured data help both.
- Keep pages current. Refreshing older content kept our rankings steady through algorithm updates, and it’s what kept us cited.
- Build trust off your own site. Bylines, partner listings and third-party mentions are part of the content strategy now.
- Measure it. We set up GA4 reports for AI referrals and watched them grow from almost nothing to more than 9,500 visits.
AI visibility also moves fast. Citations can rise and fall within the same year, often for reasons outside your control. That’s why I don’t chase it as its own channel. I build content that deserves to be cited and track whether it is.
The part that hasn’t changed
The pages AI tools trust most are the ones written for a real person with a real problem: specific, honest about trade-offs and useful from the first line. That was good content before AI search. It’s still good content now. The difference is that the shortcuts stopped working.
Keep reading
One content engine, every channel
How I built an in-house content engine that feeds sales, support, lifecycle and international marketing.
Read next Book a call Email me