Nemanja Milenkovic

18 of 200 pricing and feature claims contradicted official sources. We checked the answers against named plans, billing terms and documented limits.

We asked matched English and Japanese buying questions through two API configurations. Name overlap was 0–11.1% in the eight groups we could compare.

We inspected robots.txt on 500 corporate domains. Only 302 files could be classified without uncertainty; search and training agents often had different rules.

521 hostnames appeared in our Gemini API sample. Vendor sites led the count, sometimes linked beneath a competitor’s recommendation.

Teamwork.com appeared in 28 of 30 answers, then 2 of 30 a week later. We repeated three fixed software-buying questions and kept every answer.

Caso is a contractor in Belgrade. Our first 18 measurements said no AI assistant ever named them. The questions were wrong, not the company. Here is how to tell the difference.

A checker that refuses to call a 200 response proof, and a generator that builds a file following the proposal. Both free, no signup. Neither one will improve your rankings, and we will say so on the page.

In our S&P 500 audit, 110 companies returned a successful response at /llms.txt. Only 71 served an actual file. Here is how a status code fools a checker, and what to test instead.

We checked all 500 companies in the S&P 500. Seventy-one had a detected llms.txt. Only 54 had both a title and a Markdown link. Here is what the responses revealed.

In our stronger AI visibility run, assistants repeatedly pointed toward review sites, vendor pages, and SEO publications. That source layer may decide how buyers hear about your brand.

In a stronger AI visibility sample, Ahrefs edged Semrush by score. The more useful finding was how clearly assistants separated their jobs.

We ran a stronger AI visibility test across 10 software brands. The uncomfortable finding: assistants repeat the same pricing, setup, customization, and trust complaints buyers already worry about.

AI assistants framed OpenAI and Perplexity very differently. One looked like a broad AI platform. The other looked like an answer engine for cited research.

AI assistants repeatedly place brands next to the same competitors. That co-mention pattern is a practical map for comparison pages, buyer guides, and GEO strategy.

Early attention from AI search and content can look promising while conversion stays weak. The fix is usually the funnel, not more traffic.

A stronger CRM visibility sample found Salesforce, Close, HubSpot, Pipedrive, and Attio all scoring well. The useful signal was not the leaderboard. It was the buyer story attached to each brand.

As AI answers absorb more discovery traffic, email becomes a more important owned audience layer. That only works if the mail actually reaches the inbox.

Most growth problems are not traffic problems. A plain look at the free-tool-to-paid loop, the four conversion rates that matter, and why paid ads should wait until the funnel holds.

Search engines and AI assistants are different products that often reward the same work. A practical guide to building one content system that earns rankings and citations at once.

A plain guide to email deliverability for small senders. What SPF, DKIM, and DMARC actually do, why new domains get filtered, and the fixes that move you back to the inbox.

An llms.txt is not a one time chore. Sites change, links rot, deploys break things. Treat it like uptime, and connect it to whether AI actually recommends you.

Very few sites publish an llms.txt, and most that do get it wrong. Here is why being early is cheap leverage, and how to write one that actually works.

What llms.txt proposes, what belongs in the file, and why publishing one is not evidence of better AI visibility or search rankings.

For years the goal of online visibility was the click. As more people get answers straight from AI assistants, the new currency is being cited inside the answer.

AI assistants do not pull names out of thin air. Here is a plain explanation of what actually drives which brands get mentioned in an answer.

B2B buyers research carefully and ask detailed questions. That is exactly the kind of work AI assistants are good at, which is why mentions matter so much in B2B.

Generative Engine Optimization is real, but most GEO advice is recycled SEO. What works today, what the platforms actually document, and where this is headed.