Make the record provable
Answer engines cite sources they can verify. The firm builds your entity graph, claims your profiles, aligns your organization data, and ships proof pages that hold up when a model checks who you are.
Suede Visibility · Credibility, reputation, AI search
When buyers ask AI, be the brand in the answer.
Suede Visibility is the marketing firm from Suede Labs for the questions buyers now ask engines instead of search bars. We ship the structure, proof, and authority work that makes ChatGPT, Perplexity, and Google AI Overviews cite you, and we report share of voice engine by engine. The firm holds five client cohorts at a time. A person reviews every application.
The offerFive cohorts, hard capA person reviews every applicationFixed query set, tracked per engineSchema and proof work shippedNo guaranteed citations
The work
Three outcomes, one desk
Credibility, reputation, and visibility are usually sold as three engagements by three vendors. In AI search they are one problem: what an engine can verify about you decides what it says about you, and what it says decides whether you appear.
Audit what the engines say about you
Ask each engine about your brand by name and you get an answer today, good or bad. The firm baselines that answer, fixes the sources it draws from, and tracks how it moves.
Appear where your category is asked
When a buyer asks for the best option in your category, the answer names someone. The firm works a fixed set of those buyer prompts and reports your share of them, engine by engine.
Definitions
SEO, AEO, and GEO are different jobs
SEO earns a ranked link a person clicks. AEO wins the direct answer an engine extracts. GEO gets your brand named and cited inside the response an AI writes. The firm works all three, and reports them separately, because a brand can win one surface and be invisible on another.
| Discipline | The surface | What winning looks like |
|---|---|---|
| SEO | The ranked results page | Your page holds a position a person can click. |
| AEO | The direct answer box | Your content is the answer an engine extracts and shows. |
| GEO | The generated response | Your brand is named and cited inside the answer an AI writes. |
The method
Four phases, each with an artifact
Every phase ends in something you can hold: a dated ledger, a deployed page, a checklist of credentials you own. If a month produces no artifact, the month failed and the ledger says so.
- 01Baseline
We fix a query set of real buyer prompts for your category, then capture timestamped evidence of which brands each engine names today.
Artifact: the baseline ledger, dated, yours to keep. - 02Structure
We ship the structural work on your site: schema, entity markup, literal-question headers, and answer-shaped pages engines can extract.
Artifact: deployed markup and pages on your domain. - 03Authority
We build the proof the engines look for: consistent third-party surfaces, claimed profiles, and citable reference material, all owned by you.
Artifact: an authority checklist with every credential under your email. - 04Ledger
Every month we re-run the query set and report your share of voice per engine against the baseline, with the misses listed next to the wins.
Artifact: the monthly ledger, wins and misses both.
ChatGPTBaseline engine
PerplexityBaseline engine
AI OverviewsBaseline engine
The cap
Why five cohorts, and not fifty
Every cohort gets founder-level review of the work that ships under its name: the query set, the schema, the proof pages, the monthly ledger. Five is the ceiling where that attention holds. Past it, review turns into account management and the quality claim on this page stops being true.
When the board is full, it is full. Applications stay open and queue in order, and the next opening goes to the front of the line. The board above is the public record of the roster, set by hand, dated, and honest about the fact that a new firm starts with five open seats.
The firm runs its own checks on the rails this site documents. The node catalog behind the monitoring flows is public at /docs/nodes, and the agent directory this studio publishes is at /agents. Nothing in the method depends on tooling you cannot inspect.
Useful limits
What the firm will not do
- No guaranteed citations.Nobody controls what a model answers. The firm improves your odds through structure, proof, and authority, and reports the result honestly. Anyone promising placement is selling something else.
- No fake proof.No purchased reviews, no invented testimonials, no borrowed logos. This practice is new and says so. The founding cohorts become the case studies.
- No hostage accounts.Every profile, property, and schema deployment is claimed under your domain and your email. If we part ways, you keep everything.
- No manufactured urgency.The board on this page fills by hand and empties by hand. There are no countdown timers, no seat tickers, and no deadline that quietly resets.
Apply
Apply for a cohort seat
The application is an email, not a form. Five things, in your own words. A person reads it and replies within one business day with either a review call or a straight no.
No payment is collected with an application. Pricing and terms are quoted in the cohort proposal after review. Questions before applying go to the same address: support@suedeai.ai.
- Company and domain
Who you are and the site the work would land on.
- Three buyer prompts
The three questions you most want to be the answer for. Real phrasing beats keywords.
- Current standing
Links that represent you today: press, docs, profiles, anything an engine might already be reading.
- The trigger
What made this urgent. A lost deal, a wrong answer about your brand, a competitor cited in your place.
- Timeline
When you want the baseline captured and any date the engagement must respect.
FAQ
Before you apply
How long until AI engines cite my brand?
Structural work lands in the first weeks: schema, entity cleanup, and answer-shaped pages. Citation movement is slower and typically takes months, because engines re-crawl and re-weigh sources on their own schedule. The monthly ledger shows movement or its absence honestly, so you are never guessing.
Do you guarantee citations or placement?
No. Nobody controls what a model answers, and a guarantee in this category is a red flag rather than a promise. The firm commits to the work and the reporting: a dated baseline, shipped structure and proof, and a monthly share-of-voice ledger per engine against a fixed query set.
Which AI engines does the firm track?
The baseline set is ChatGPT, Perplexity, and Google AI Overviews. Gemini, Claude, and Copilot are added when your buyers are there. Share of voice is reported per engine because the same brand can be strong in one engine and absent from another.
What does an engagement cost?
Pricing is quoted in the cohort proposal after a person reviews your application. No payment is collected with the application, and the free scan costs nothing. The quote names the retainer, the term, and exactly what ships each month.
Why does the firm only take five cohorts?
Every cohort gets founder-level review of the work that ships under its name. Five is the ceiling where that attention holds without diluting into account management. The cap is an operating constraint, not a marketing device, and the board on this page is its public record.
What happens if the board is full when I apply?
Applications stay open and queue in order. When an engagement wraps, the next application in line gets the review. The board statuses are set by hand and dated, so what you see on this page is the actual state of the roster.