In 2025, in my role as head of AI at BBN, I wrote a survival guide for B2B marketers on what was then a shiny new acronym: GEO, Generative Engine Optimization. It had everything a good survival guide needs. An acronym soup (SEO, AIO, AEO, GEO, SXO), a four-pillar framework, a checklist, even a scoring exercise for your next team meeting. I was quite proud of it.
Rereading it now in mid 2026, from the perspective of the BBN AI Studio, most of it holds up. The core prediction certainly did. Your prospect opens an AI chat, asks about automation solutions or a specific feature on an agricultural machine, and gets a detailed answer without ever visiting your carefully optimized website. That scenario stopped being a ‘picture this’ and became a reality.
But something else happened in the meantime, and it’s the reason I’m rewriting the piece instead of just reposting it. An entire industry sprang up promising to fix this for you. Hand over your brand name, the pitch goes, and our tool will make AI recommend you. Set it, forget it, watch the citations roll in.
We don’t believe that. And I say this as someone who builds AI tools for a living.
Why there is no autopilot for AI visibility
Here’s the problem with every ‘we’ll fix your AI visibility’ promise. Your visibility is not a generic problem, so it can’t have a generic solution.
Your brand is unique. Its authority, its niche, the questions your buyers actually ask. Your competitors change their activities every single day, which means the answer space AI draws from changes with them. And the AI platforms themselves keep shifting how they select and cite sources. A tool that promises to automate all of that is automating guesswork.
What made me even more skeptical is what we learned from watching how AI systems actually cite. The citation pool is consolidating. AI answers increasingly draw on fewer, more authoritative sources. The winners aren’t the brands with the cleverest optimization hack. They’re the ones whose content contains something the model can’t get anywhere else: first-hand data, real experience, a point of view.
You cannot outsource having something to say.
Since AI arrived, we’ve thought as hard as anyone in marketing about the visibility implications of all this. Our conclusion, after two years of testing: there is only so much AI out of the box can do for you, and there is nothing a hype tool can do that the fundamentals done well can’t do better.
The good news: the fundamentals are teachable
Now for the part of my old article that aged well. The fundamentals of getting found, by Google and by AI, haven’t changed as much as the tool vendors want you to think. In fact, my strongest conviction in 2026 is that GEO and SEO are one discipline. Optimizing for AI answers is search strategy applied to AI.
Structure your content so it can be understood and extracted. Answer the questions buyers actually ask, in the way they ask them. Show your credentials. Put your expertise where machines and humans can read it. If your best thinking is locked in a gated PDF, AI simply doesn’t know you have it. Lead with the answer instead of burying it. Add something the top ten results don’t already say.
None of that is magic. All of it is work. It’s the kind of unglamorous, repeatable work that senior search consultants have been quietly doing for years. And that’s exactly why we could build it into a skill in the BBN AI Studio, our platform.
Our skill scores a draft across seven dimensions. The one it weights highest is differentiation and information gain: does the page say anything the rest of your industry isn’t already saying?
Our GEO and SEO skill doesn’t promise citations. It reviews your drafts the way a senior consultant would. It scores the content across seven dimensions, from structure and AI extractability to authority signals to the one we weight highest: differentiation and information gain. Then it hands you a prioritized list of recommendations, each tagged for search, for AI visibility, or both, ranked by effort versus impact. It’s opinionated. It will tell you to add a TLDR and why. What it won’t do is claim it fixed your visibility while you slept.
The reality test
Here’s what convinced me this belongs in daily practice rather than in a strategy deck. For one of our clients, I took a draft of a landing page and ran it through the skill on our platform.
It came back with a long list of concrete improvements. Restructure sections so an AI can extract clean answers. Cover the conversational questions a plant engineer would actually ask instead of the keyword the page was built around. Surface the proof points and credentials that were sitting three scrolls down. Lead with a direct answer. None of the recommendations were exotic. Every one of them made the page more viable for AI search. Not coincidentally, the recommendations made the page better for the human reading it, too. The whole review took minutes, and the quality bar was one I’d defend in front of any search consultant.
SEO isn’t dead. It just got a software update that nobody has the patch notes for.
That’s the honest division of labor. The skill does the fundamentals at a senior level, every time, inside the AI tools your team already uses. Your team keeps the part no tool can take over: the strategy, the first-party data, and the judgement about what your brand should be known for. As I put it in the original survival guide, and it’s truer now than when I wrote it: SEO isn’t dead. It just got a software update that nobody has the patch notes for.
Where to start
If your AI visibility worries you, don’t buy an autopilot. Take one important page and review it against the fundamentals. Can an AI extract a clear answer from it? Does it cover real buyer questions? Does it prove who’s behind it? Does it say anything the rest of your industry isn’t already saying? Then fix what you find, and repeat with the next page.
That loop is exactly what our skill runs, and it’s the least hyped, most reliable path to being found in 2026. The fundamentals, done well, at scale. Everything else is somebody selling you a shortcut on a road that doesn’t have one.