Case Study

The Coaching Site AI Assistants Now Recommend First

An expensive agency website that produced zero leads. We rebuilt it around AI-answer-ready landing pages. Today it tops AI recommendations for many coaching questions in its market, and the client reports roughly 10x the clients.

claw.mobile Editorial·7 min read·August 21, 2026

Quick Answer: What this case study shows

A professional coaching practice in Portugal paid an agency for a polished website that generated zero leads. We rebuilt it as a set of AI-answer-ready landing pages: one page per real buyer question, direct answers up top, structured data throughout, dated content published on a schedule. The site now tops AI recommendations for many coaching-related questions in its market, and the client reports their client base grew roughly 10x.

What was broken

The client is a professional coaching practice in Portugal. We are keeping the brand anonymous at their request, and every claim here is one we would repeat with them in the room.

When they came to us, they had already done the thing conventional wisdom says to do: hired a proper agency and paid proper agency money for a website. The site was genuinely pretty. It was also producing zero leads. Not few. Zero. Every client the practice had came from referrals and word of mouth; the expensive site was a business card nobody read.

The audit took an afternoon and the diagnosis was the same one we write about in our pillar piece on AI invisibility. The site was five pages of brand language: “transformative journeys,” “unlock your potential.” No page answered a question a prospective client actually asks. No pricing signals. No structured data at all. No dates on anything. Beautiful, and unreadable to machines.

Meanwhile the buyers had moved. People choosing a coach ask assistants things like whether coaching is worth it, what it costs, how to pick someone credible, who works with people in their situation. We ran those questions through ChatGPT, Claude and Perplexity. The practice appeared in none of the answers. The engines had literally nothing about them to work with.

What we changed

We did not redesign the brand. We rebuilt the site’s information architecture around a simple rule: every important page must be the best available answer to one real question a buyer asks. Four changes carried the result.

1. Answerable landing pages instead of brand pages

We mapped the questions people in the practice’s market actually ask before hiring a coach and built a landing page for each: question as the title, direct answer in the first paragraph, honest detail below, including the questions competitors avoid, like cost and what coaching cannot do. The vague brand copy moved aside; the answers moved in.

2. Structured data on every page that matters

Organization and Service schema describing what the practice is and offers, FAQPage markup on every landing page, all in validated JSON-LD. The site went from a wall of divs to something a retrieval crawler can parse in one pass.

3. Dated, methodical content on a schedule

Engines favor sources that look alive and specific: visible dates, regular publishing, concrete detail over adjectives. We set up a steady cadence of dated articles answering the longer-tail questions, in the practice’s own voice, and kept it running rather than launching once and going quiet.

4. Nothing hidden from machines

Every word server-rendered, no content locked behind click-to-expand JavaScript, clean headings, fast pages. LLMs do not run your scripts; what they cannot read does not exist. The rule was boring and absolute.

What happened

The visibility side we can verify ourselves, because we measure it the same way we measure our own market: run the buyer questions through the engines, log the names that come back. Today, the practice tops the AI recommendations for many coaching-related questions in its market. Ask the assistants the questions their buyers ask, and this practice is the name that keeps coming back, on questions where a year ago it appeared in nothing.

The business side belongs to the client, so we state it the way they state it: since the rebuild, their client base has grown roughly 10x. That is their reported number, not a screenshot from their analytics, and we are not going to dress it up with charts we did not produce. From zero leads to a practice that grew by an order of magnitude, with the website as the thing that changed.

What we like most about this result is how unmagical it is. No tricks, no loopholes, nothing that gets burned by the next model update. The site earned its position by being the most readable, most direct, most honestly informative source in its niche, which is exactly what the engines are trying to find. The same playbook transfers: we have written up how it applies to local businesses and B2B SaaS.

Want this for your business?

Two ways we do this for clients. AI-ready landing pages: we build the answerable, structured, machine-readable pages described above for your market, as part of our development services. And the AI Visibility Report ($1,990): before building anything, we run 40-50 of your buyers’ questions across ChatGPT, Claude and Perplexity and hand you your current share of voice, the sources the engines cite in your market, and a prioritized plan. Ongoing tracking is $490/month.

Start with the audit →

Frequently Asked Questions

Why is the client anonymous?
Because it is their business and their call. The practice competes in a small market where being publicly dissected as a case study has real downsides, so we describe it as a professional coaching practice in Portugal and leave the brand out. The owner can choose to be named later. Every claim in the article is one we can stand behind with the client in the room.
Where does the 10x number come from?
From the client. It is their reported growth in client base since the rebuilt site went live, not a number we pulled from their analytics, and we publish no screenshots or dashboards from their accounts. What we can verify directly is the visibility side: we run coaching-related buyer questions through ChatGPT, Claude and Perplexity ourselves and watch which names come back.
What is an AI-answer-ready landing page?
A page built so an AI assistant can lift the answer straight out of it: one real buyer question as the title, a direct answer in the first paragraph, the reasoning and specifics below, a visible date, FAQ markup in JSON-LD, and no content hidden behind JavaScript. It reads well to a human and parses cleanly for a machine. We build these as a productized service.
Would this work for a business that is not coaching?
The mechanics are not coaching-specific. Any business whose buyers ask an assistant questions before hiring someone, which now includes dentists, lawyers, agencies, and most of B2B software, is subject to the same dynamics: the engines recommend whoever the readable, trustworthy sources describe. The starting move for any niche is the same: run your buyers’ questions through the engines and see who they name.
How long did the turnaround take?
The rebuild itself took weeks, not months. The visibility shift came gradually after launch as the new pages were indexed and picked up: engines with live retrieval reacted first, and coverage broadened from there. This was not overnight, and anyone promising overnight is selling something else.

Need a website or bot built?

Fixed pricing from $999. Free mockup in 48h. You own the code.

See pricing

Get the Vibe Coding Cheat Sheet

Best tool for every use case + pricing + pro tips. One page, zero fluff. Plus weekly updates on new tools.