This is Part 2 of our AI SEO, AEO and GEO experiment. In Part 1, we documented why 99francs.agency was effectively invisible for the commercial categories we wanted to own, then changed the site architecture, answer blocks, schema, internal links and machine-readable discovery paths. Here is what happened in the first 31 days.
The experiment, in one sentence
We treated our own site as the product: make the offer, proof, prices, comparisons and answers easier for Google, AI Overviews and LLMs to understand — then measure whether that wider understanding creates qualified demand.
The 31-day result
| Metric | Before: May 8–June 7 | After: June 8–July 8 | Change |
|---|---|---|---|
| Organic clicks | 3 | 11 | +267% |
| Google Search impressions | 59 | 479 | +712% |
| Queries with impressions | 1 | 73 | +72 queries |
| Qualified leads | — | 3 | 3 attributed leads |
| Orders | — | 2 | 2 attributed orders |
| LLM-sourced leads | — | 1 | 1 attributed lead |
The periods are deliberately equal: 31 days before the June 7 baseline and 31 days after it. We do not include average position in this table. Before the work, Search Console had meaningful data for essentially one query; after it, Google was testing the site across 73 queries. A blended position would confuse wider discovery with a ranking decline.
What Google started testing us for
The most useful movement was not a vanity keyword. Google began showing the site for commercial website and development questions — the kind of searches that can lead to a project rather than a casual read.
- website design nz cost — 39 impressions
- website design cost — 24 impressions
- web development agencies — 15 impressions
- how much does a website cost nz — 15 impressions
- best webflow agency nz — 14 impressions
- website design cost nz — 12 impressions
- web development companies in new zealand — 11 impressions
These are early positions, not a claim that we now rank first. But they show Google has started associating the site with commercial intent we were previously absent from. The New Zealand pages, including website development for New Zealand and our NZ website-cost guide, helped give those queries a relevant destination.
The business result matters more than the graph
In the same period, the work generated three qualified leads and two orders. One of the three leads told us they found 99 Francs through an LLM. That does not prove that any single schema block, llms.txt file or article caused the conversion. It does prove that answer-engine discovery has become a real acquisition path for us — not just a conference-slide idea.
The useful question is not ‘did an LLM cite us?’ It is ‘can the right buyer understand, trust and contact us after an AI answer introduces the company?’
Alexander Vinokurov, 99 Francs
What we changed
The work was not one technical trick. It was a connected content and conversion system — the kind of scope we deliver through AI SEO & AEO when a company has strong work but an unclear search presence.
- Clarified the company entity across canonical URLs, Organization schema, service pages and proof.
- Added direct answers, FAQ coverage and structured data where users and machines needed a concise explanation.
- Published buyer guides, comparison pages and pricing context around real decision-stage questions.
- Connected services, industry pages, case studies and articles with intentional internal links.
- Published llms.txt and kept sitemap, robots and canonical discovery paths consistent.
- Made commercial pages clearer about scope, proof, pricing and next actions — so new visibility had somewhere useful to go.
What this experiment does not prove
Thirty-one days is too short to claim a durable ranking win, and several changes went live together. We cannot isolate the causal effect of schema, internal linking, new articles or llms.txt. We also cannot use Search Console to measure LLM mentions. The LLM lead is a verified attribution from the lead itself, not an estimate of a platform-wide share.
That constraint is part of the point. AI SEO, AEO and GEO should not be sold as a magic ranking switch. They are practical work to make a company easier to understand, easier to cite and easier to choose — then measured against search data and actual pipeline.
What we will measure next
- Whether the new commercial queries move from testing impressions to stable clicks and top-10 positions.
- Which service and buyer-guide pages create the next qualified contacts.
- Whether more leads self-report LLM discovery, and from which answer engines.
- Whether third-party mentions and comparison coverage improve the frequency and quality of AI answers about the studio.
If your site is good but hard to discover
The lesson from our first month is simple: more content is not the objective. A site needs a clear entity, commercial pages that answer buyer questions, proof, a crawlable internal structure and a path from discovery to contact. If your company has the work but search and AI systems struggle to describe it, see our AI SEO & AEO service. We can audit the gaps, map the pages that matter and implement the system with your team.
For the original baseline, implementation checklist and rationale, read Part 1: AI SEO, AEO and GEO for Startups.
Frequently asked questions.
Need this done instead of just read about it?
99 Francs is a subscription-based design studio: one flat monthly rate, unlimited requests, first delivery in 1–2 days. Start with pricing or book a free intro call.
