Maybe someone suggested a Signals Audit, or you saw it on our site and wondered what you'd actually be buying. Fair question. "Audit" can mean anything from a quick checklist to a 60-page report nobody reads.
Here's what it is in plain words, who it's for, what you get, and how to prepare, plus a 20-minute version you can run on yourself today.
What is a Signals Audit?
A Signals Audit is the entry-point service at Seats at the Table (SATT). It reviews what your business communicates when you're not in the room: to a potential client reading your website late at night, and to AI tools like ChatGPT, Perplexity, and Google's AI answers.
It covers how clearly you're positioned, what your website says, how easily your content can be found, how visible you are in AI search, whether you can be told apart from businesses with similar names, and the gap between what you say you do and what prospects actually experience. It's organized around the SATT Framework, the four-part method behind all our work.
What does each part of the audit look at?
The framework has four areas. They're checked in a deliberate order, because each one depends on the one before it.
Signals: what your business says about itself. Every page, profile, and listing makes a claim about what you do. Signals asks whether those claims are clear and agree with each other. After reading your homepage, could someone say what you do and who it's for in one sentence? Does your LinkedIn describe the same business? Is an old description still floating around?
Access: whether people and AI tools can find and read it. Clear words don't help if nothing can reach them. Access looks at whether search and AI tools are allowed to read your site, whether your important pages can be found, and whether your content answers buyers' real questions in a form that can be quoted.
Touchpoints: whether the path from "interested" to "hired" holds. The contact form, the booking link, the first reply, how easy it is to tell whether you're a fit. One broken step quietly wastes everything before it.
Trust: whether you've earned a recommendation. Specific proof with real names on it, a consistent description of you across the web, and other people mentioning you. Trust is what makes a person, or an AI tool, comfortable saying your name.
The order matters. Trust work, like publishing more, amplifies whatever your Signals already say. If the message is unclear, more of it makes you more visibly unclear.
Who is a Signals Audit for?
It's built for founders and small teams, typically one to ten people, who know their work is good but suspect the market isn't getting the message. Maybe your referrals have slowed, or you changed direction and aren't sure anyone noticed. It's the starting point we recommend for most new clients, alongside a paid strategy session.
It isn't designed for enterprise companies, for other agencies, or for anyone choosing mainly on price.
What will I get at the end?
A prioritized action plan, not a report to file away. Findings are ranked by how much they're likely to recover and how much effort they take, so your first week of work is the highest-return week rather than the easiest one.
You'll know what your business is communicating today, where that message breaks down, and the order to fix it in. Many high-value fixes are unglamorous, like making your description match everywhere. Unglamorous usually means fast. What it won't do is promise you a place in ChatGPT's answers or a search ranking. Nobody controls those.
How should I prepare for a Signals Audit?
Don't clean anything up first; the point is to see things as they are. A little preparation helps:
- Your own one-sentence description of what you do and who it's for, written before you look at your website.
- A list of where you show up: website, LinkedIn, Google Business Profile, directories, podcast or guest appearances.
- The questions your best clients asked before they hired you, in their words.
- Who you want more of, and who you'd like fewer of.
- Anything you've changed recently, like a new name, a new niche, or a retired service.
Can I do a 20-minute version myself?
Yes. It's rougher, but it will show you whether something's off. Use a private browser window.
- Signals, 4 minutes. Write one sentence: what you do, for whom, and the result. Then read the top of your homepage, your LinkedIn headline, and your Google Business Profile. Do all three describe the business in your sentence?
- Signals, 3 minutes. Search your business name. Note anyone else using a similar name, and any old description or retired service still showing up.
- Access, 5 minutes. Ask ChatGPT or Perplexity "what does [your business] do?" and one question a buyer would ask, like "who helps [your client type] with [your problem]?" Write down whether you're named and how you're described.
- Touchpoints, 5 minutes. Send an inquiry through your own contact form from a different email address. Did it arrive? Did the confirmation make sense? Note how long a real reply would take.
- Trust, 3 minutes. Count the proof a stranger could check: reviews, client stories with real names, articles or interviews that mention you. Is any of it from the last year?
If step one turned up a mismatch or step four turned up a problem, fix those first. Our posts on why ChatGPT doesn't recommend some businesses and what to put on your website for AI search cover the next layer.
What shouldn't I do before fixing the basics?
- Publish more content to fix a visibility problem. Volume amplifies an unclear message.
- Pay for any service that guarantees AI placement or rankings.
- Treat one technical file as the fix. Google says you don't need special AI text files or markup to appear in its AI features; the fundamentals still do the work.
- Change everything at once. Without a baseline, you won't know what worked.
What happened when we ran it on ourselves?
We ran the SATT Framework on Seats at the Table before offering it to anyone. We had just repositioned and thought we were done. We weren't. Search results still showed our old homepage title and a positioning we'd retired. The file we'd written specifically to tell AI tools who we are described the old business. Our own site called one part of the framework by two different names. And several unrelated organizations share a name close to ours, including a UK consultancy, two podcasts, and a networking group in Asia, which makes it easy for AI tools to mix us up.
Then, in August 2026, we found that our own website contact form had never worked. A placeholder setting meant inquiries silently went nowhere. We fixed it on August 14. None of it showed up on any dashboard we owned. That's why the audit exists.