Signal Faible: crawlable, but not yet easy to name

Creative studio, Boulogne-Billancourt, France. AI Visibility Audit, with the visibility work in progress. Audit run 2 September 2026; client work began 3 September 2026.

Where it stands

Every case is scored on the same three steps. Read: can AI systems reach and read the pages that matter? Trust: do the business’s facts match everywhere they appear? Recommend: do outside sources back the business up where AI looks?

Step	Found on 2 September	Now	Status
Read	Readable, but a duplicate homepage competed with the real one and held the studio’s story	Duplicate gone; the studio story still needs its own page	In progress
Trust	13th for its own name; a thin business record; no analytics	The business record still carries four fields (name, address, image, web address); analytics still to install	Open
Recommend	On none of the directories AI answers cite for the category; no Google Business Profile found	Profiles and directory listings are in the plan below	Not started

One week in, the duplicate homepage is gone and a contact form now sits next to the email link (both checked 10 September 2026).

What this proves

This is the in-progress case.

The site was not broken. It loaded, redirected correctly, served its text to crawlers and could be read by AI systems. That was not enough.

The business still had the harder problem: machines could read the site, but they could not confidently understand which “Signal Faible” this was, why the studio should be trusted, or where else on the web that claim was confirmed.

That is the work now in motion: make the entity unmistakable, build the sources assistants cross-check, install measurement, and turn a business-card site into something a buyer can act on.

The context

Signal Faible is a French creative studio for product communication: strategy, art direction, photography, video, CGI, motion design and post-production. The studio is based in Boulogne-Billancourt and works for brands in beauty, luxury and product-led categories.

The live site was small by design. Three real pages. Beautiful visual work. Almost no structure for a machine or a buyer to follow.

The client was already seeing some AI-sourced attention, but not from the right people. Internship and collaboration requests were coming in. Client inquiries were not.

Finding 1. The name belongs to somebody else

“Signal faible” is not only a studio name. In French it is an established business concept, the weak signal: the faint early indicator of change. It is also the phrase a television or wifi error message uses when reception drops.

That collision controls the search result.

For the query “signal faible”, Google’s AI answer was about the concept, not the studio. The organic result set was the same story: Wikipedia, a government service, a dictionary, a TV support thread and other concept-led pages.

The studio ranked 13th for its own name.

So the job is not to “win” the phrase. That would be the wrong fight. The job is to make the business unmistakable: Signal Faible, the creative studio, with structured data, third-party profiles and consistent qualifiers attached to the name.

Finding 2. Classic search has a low ceiling

The search demand is thin in the places that matter. The two commercial terms with verified demand were both shrinking:

Search term	Monthly searches	A year earlier
agence de communication luxe paris	90	140
agence de communication boulogne-billancourt	50	110

Even a perfect ranking on both would be a small prize: about 140 searches a month, and falling. That changed the recommendation.

Do not sell this client a blog programme built around low-volume keywords. The commercial work is entity visibility, portfolio proof, third-party presence and a lead path for the people already looking through AI, referrals and social proof.

Finding 3. The sources AI would check were missing

On the main commercial query, the visible sources were directories and shortlist pages: Sortlist, PagesJaunes, La Fabrique du Net, Mon Expert Digital and PagesPro.

Signal Faible was not on them.

I also found no Google Business Profile in the map results, no company LinkedIn page, no public review surface, and no linked social proof from the homepage HTML. The only third-party sources confirming the company existed were legal registries.

When somebody asks an assistant for a studio or agency in a place, the assistant checks the same kind of corroboration. A business that appears on none of those sources is hard to recommend, even when the work is good.

Finding 4. The site had proof, but not in a usable shape

The strongest credibility asset was Julien Celice’s story: fashion, beauty and spirits work, then co-founding Signal Faible in 2022.

That story existed on a leftover duplicate homepage.

The audit found /copie-de-accueil live, indexable, canonical to itself, listed in the sitemap and carrying the same title and description as the homepage. In practice, the best credibility content on the site was hidden inside a duplicate page competing with the homepage.

The fix is simple in concept: turn it into a real studio page, with its own address, title, description, heading and internal links.

Finding 5. Measurement did not exist yet

There was no GA4 installed on the site.

The live tag was a Google Ads conversion tag, not an analytics setup. There was also no form, no booking route and no conversion event to record. The site had one real lead path: a mailto link.

So the first measurement job is not reporting. It is creating something worth measuring: a real lead path, GA4, Search Console, Bing Webmaster Tools, and a basic conversion event.

What is being improved now

The work is underway, not finished.

The first sequence is deliberately practical:

1. Decide whether the live site stays on Wix or moves to the rebuilt custom version.
2. Fix the duplicate homepage and give the studio story its own page.
3. Rewrite the English metadata so the English page does not ship French search copy.
4. Create or claim the Google Business Profile.
5. Complete the structured data: legal name, company number, founder, address, services, telephone, price range and linked profiles.
6. Create the company LinkedIn page and the directory profiles machines already use.
7. Add portfolio pages so the site has work a buyer can inspect and an assistant can cite.
8. Install analytics and a real lead path.

That is the point of this case study for a website agency. The website is part of the answer, but not the whole answer. The commercial layer is deciding what proof belongs on the site, what proof must exist outside it, and how a buyer moves from “I found you” to “I want to talk.”

The current state

No uplift is claimed here yet.

The baseline is the value:

Baseline check	Result
Real pages found	3
Ranking keywords in France	1
Estimated organic visits per month	About 2
Position for the brand phrase	13
Top-90 appearances for main commercial query	0
Followed backlinks	9 of 37
Unique words across the site	1,192
GA4	Not installed
Google Business Profile	None found

Now the work has a clean starting line.

Evidence

---

Sources. Signal Faible visibility audit, run 2 September 2026 against the live site; direct requests to signalfaible.com including crawler-agent tests; SE Ranking for rankings, search volume and link data; live French search-result checks; French company registry record 914253802; Signal Faible discovery call, 3 September 2026. Search Console, true indexed page count and finished performance impact require access and are not claimed here.