The best thing you can do for most clients right now has nothing to do with rewriting their homepage. It's leaving it alone — and adding something new instead.
A well-built llms.txt file is the cleanest, lowest-risk AEO move available today. It doesn't touch existing pages. It doesn't risk Google rankings. And it gives AI assistants a direct line to the trust signals they weight most heavily when deciding whether to recommend a business.
That's the honey. Let's talk about what goes in the jar.
Every reputable SEO operator knows the same rule: don't overhaul a site's pages because an optimization trend told you to. Unnecessary edits create ranking risk — and right now, a lot of AEO advice is pushing exactly that kind of churn on client sites.
The smarter play is additive, not destructive. The llms.txt file lives at the root of a domain, it's publicly accessible, and it's written specifically for AI crawlers. It doesn't interfere with anything Google is already indexing. Your clients get AEO coverage without you touching a single title tag.
That's a product your clients will actually thank you for.
Among the AI Ranking Factors we test and track at viizable, one of the most consistently weighted signals is entity verification — the AI's confidence that a business is real, stable, and rooted in a specific place and community.
When an AI assistant answers "who's the best HVAC company in [city]," it's not just pulling from the business's website. It's cross-referencing. It wants corroboration. A claim on a site is just a claim. A claim on a site that matches something it can verify elsewhere — that's evidence. And evidence is what makes AI confident enough to give a recommendation.
This is the whole mechanic behind AI honey: structured, linked proof that your client's business exists, is established, and is recognized in its community.
Think of the llms.txt file as a trust briefing for AI crawlers. It's not a sales page. It's not a bio. It's a curated index of the things that make a business verifiable — the details the public might never think to look for, but that AI weights heavily.
Some of the highest-value entries to include:
A note worth passing to clients: this file is publicly accessible, so nothing genuinely sensitive should go in it. But most of the trust signals that matter to AI are things businesses are already proud of — they just haven't organized them anywhere AI can actually read efficiently.
Here's the mechanic in plain terms: AI assistants are trying to minimize the risk of giving a bad recommendation. The more a business's claims are echoed by independent, verifiable sources, the more confident the AI becomes.
When your client's llms.txt file points to a government license record, a newspaper feature, and a trade association profile — all independently confirming the same business, in the same city, doing the same work — that's not a website anymore. That's an entity. And entities get recommended. Unverified pages get overlooked.
This is especially true for local queries. When someone asks an AI assistant for a recommendation in their neighborhood, the businesses that come back aren't just the ones with the most content — they're the ones the AI trusts it can stake its credibility on.
The other reason this matters for agencies: visibility into AI crawl activity. A properly structured llms.txt file isn't just a trust signal — it's an observable touchpoint. When AI crawlers come through and engage with that file, you can track it. That's reporting your clients haven't seen before, and it tells a story about AI visibility that no traditional rank tracker can show.
If you want a fast way to build these files, llmsmaker.com pulls existing site data and structures it in the order AI crawlers prioritize — and it's free to use. For the agencies reading this, it also gives you a built-in reporting hook for client conversations.
The agencies winning in AEO right now aren't the ones making the most changes. They're the ones making the right ones. A well-structured llms.txt file is one of the few moves that's genuinely additive — it improves AI visibility, creates new reporting capability, and carries zero risk to existing rankings.
That's not a trend. That's a durable play. And your clients will be asking why you didn't build it sooner.