Agent-Readiness: Can an AI Find Your Business? We Tested 62 Atlanta Sites, and Most Can't

Agent-Readiness: Can an AI Find Your Business? We Tested 62 Atlanta Sites, and Most Can't

Here is a thing that is about to happen, if it hasn't already happened to you.

Someone asks their phone, "book me a facial near Buckhead this Saturday," or "find an emergency plumber that's open right now," and an AI assistant goes and does it. It doesn't browse ten blue links. It reads businesses, decides, and acts, on the person's behalf. The customer never sees your website. Their AI does.

So the question stops being "does my site look good to a person?" and becomes something new: can a machine read my business at all? We are calling that Agent-Readiness, and we decided to measure it.

We took the same 62 real Atlanta businesses from our website speed study and asked a different question of each one: not how fast it loads for a human, but whether an AI agent could actually understand it. The answer, mostly, is no.

What Agent-Readiness means (and the score)

An AI agent doesn't "look" at your page. It reads the underlying data and the raw HTML. To act on your business, it needs five basic things to be machine-readable. We scored every site 0 to 5:

1. Can it identify you? Structured data (schema.org) that says "this is a business, here's its name and type."
2. Can it tell if you're open? Your hours, in machine-readable form, not just a JPEG of a sign or text buried in a script.
3. Can it reach you? A phone number in the structured data.
4. Can it find you? An address in the structured data.
5. Can it read you at all? Your actual content present in the HTML, rather than a JavaScript-only shell that looks blank until a browser runs code an agent may not.

We deliberately did not have AI agents make real bookings, that would mean booking real slots at real businesses. We measured the foundation underneath booking: whether the business is legible to a machine in the first place. If an agent can't read your hours, it can't even consider you when someone says "who's open now."

The results: most Atlanta businesses are invisible to agents

Of the 58 sites we could analyze (four blocked our request entirely, more on that below):

• Only 28% were fully agent-ready (all five signals present).
• 43% scored a 0 or a 1, effectively invisible to an AI agent trying to act on them.
• Just 33% publish their hours in a machine-readable way. Two out of three Atlanta businesses we checked cannot answer the single most common local question an AI gets asked: "are they open?"
• 57% identify themselves with business structured data. 40% expose a machine-readable phone; 43% an address.
• 90% at least had their content in the HTML. The other 10% are JavaScript-only shells that an agent, or a plain crawler, sees as nearly empty.

Put simply: when the assistant economy arrives, most of these businesses won't be chosen against, they won't be seen.

The scoreboard, by industry

Average Agent-Readiness score out of 5:

Salons came last, and it makes sense: many run entirely on a third-party booking widget with a nearly empty website behind it. That's convenient for the owner, but to an AI agent the actual business, its hours, its services, is locked inside an embedded tool the agent often can't traverse from the page.

Restaurants scored low for a similar reason: gorgeous, image-heavy sites with the menu trapped in a PDF or an image, and the hours living in a graphic. Beautiful to a human, unreadable to a machine.

The sites that blocked us entirely

Four of the 62 refused our request outright, two returned a hard "403 Forbidden" that actively blocks automated visitors. That feels safe, but it's worth thinking about: the same wall that blocks a scraper also blocks the AI agent your customer is using to find you. In an agent-driven world, aggressively blocking bots can mean blocking business.

Test your own Agent-Readiness in two minutes

You can get a rough read yourself:

1. Go to Google's Rich Results Test, paste your URL, and see whether it detects LocalBusiness structured data with your hours, phone, and address. No detection means an agent has nothing to read.
2. In your browser, view your site with JavaScript disabled (or use "View Source" rather than "Inspect"). If the page is blank or missing your key info, so is it to an agent.

If both come back thin, you scored what most of Atlanta scored.

Why this is the SEO shift nobody's prepared for

For twenty years, "being findable" meant ranking for humans typing queries. That era isn't over, but a second layer is forming on top of it: being legible to machines that act for humans. Google's AI Overviews, ChatGPT, Perplexity, and the coming wave of booking-capable agents all read structured data first and pretty pixels never.

The businesses that win the next few years won't just have fast, good-looking sites. They'll have sites a machine can parse, identify, and act on, with hours, services, and contact details exposed as data, not decoration. Right now, in Atlanta, fewer than a third are there.

How TiltStack builds for this

Every site we build ships with the structured data an agent needs, LocalBusiness schema with machine-readable hours, services, and contact details, server-rendered content an agent can actually read, and booking paths that aren't locked behind an opaque widget. It's the same work that helps you rank today, done in a way that also keeps you visible as search becomes agentic.

Want to know your Agent-Readiness score? Send us your website and we'll run it, free, alongside the speed audit from our Atlanta website study.

FAQs

Q1: What is Agent-Readiness?
A: It's how well an AI agent, the assistants people increasingly use to book, order, and research on their behalf, can read and act on your business. It comes down to whether your key facts (identity, hours, phone, address) are machine-readable structured data, and whether your content exists in the HTML rather than only after JavaScript runs. We score it 0 to 5.

Q2: How is this different from SEO?
A: Traditional SEO optimizes for humans finding your ranked pages. Agent-Readiness optimizes for machines understanding and acting on your business. They overlap (structured data helps both), but a site can rank acceptably for humans and still be invisible to an agent if its hours and services aren't exposed as data.

Q3: How did you measure it without making bookings?
A: We fetched the raw HTML of each site (what a non-JavaScript agent or crawler sees) and checked five machine-readable signals: business structured data, hours, phone, address, and whether content was server-rendered. We did not attempt real bookings, that would mean reserving real slots at real businesses. This measures the foundation an agent needs before it can book anything.

Q4: My site is on a booking platform like Vagaro or OpenTable. Am I covered?
A: Partially, and it's a trap. Those platforms handle the booking, but if your own website is a near-empty shell that just links out to the widget, an agent landing on your site has almost nothing to read about you. You want your identity, hours, and services legible on your own page, with the booking as one clear path, not the only content.

Q5: What's the single highest-impact fix?
A: Add LocalBusiness structured data with your opening hours. Only a third of the Atlanta businesses we tested had it, and it directly answers the most common question an AI gets about a local business: "are they open right now?"