Your clients are asking if they show up in ChatGPT.
Now you can answer, under your own brand.
Semantic Scale monitors how your clients appear in ChatGPT, Gemini, Perplexity and Google AI Overviews, then turns it into white-label reports your agency sells as its own service line. Sampled properly. Explained honestly. Set up in minutes.
- New client live in under 10 minutes
- Your logo, your colors, your report
- Priced per client, not per query
The gap
“Are we showing up in AI search?” is the question every client is about to ask. Most agencies can't answer it.
Asking ChatGPT once isn't data
AI answers change from run to run, by location, by session. A screenshot of one good answer is not a measurement, and a skeptical client (or their lawyer) knows it.
Enterprise tools don't fit agencies
The monitoring platforms that exist are priced and packaged for one big brand, not for an agency running thirty small clients at a margin.
Building it in-house is a distraction
Scheduling multi-LLM queries, parsing noisy responses, and presenting results defensibly is a data-engineering project. You're an agency, not a data team.
How it works
From signup to a monitored client in one sitting.
- 1
Add the client
Brand name and aliases, domain, competitors, category, geography. Pick from a library of local-intent questions or write your own in the client's customers' words.
- 2
We sample AI answers on schedule
Every question is asked repeatedly on every surface, every cycle, on your own API keys. Failed runs retry and are logged, never silently dropped.
- 3
We detect what matters
Whether the brand was mentioned, whether its site was cited, where competitors appeared, and which sources the AI leaned on when your client was absent.
- 4
You deliver the report
Branded reports generate weekly or monthly, with a preview step so you can add commentary. Alerts fire when a brand disappears or a competitor jumps in.
What you get
Everything you need to sell AI visibility as a service. Nothing you don't.
Share of voice you can defend
Every number ships with its sample size and date range: "38% (152 of 400 answers, Aug 14–Sep 13)." Never a fake rank.
Four surfaces, one report
ChatGPT, Gemini and Perplexity through official APIs. Google AI Overviews through a licensed data provider. No scraping, no gray areas.
Competitor overlay
See who's being recommended instead of your client, cycle over cycle, so the next pitch writes itself.
Citation gaps
The domains AI cites when your client is absent, ranked. Hand your content and link team a concrete target list.
White-label everything
Your logo and colors on every report, share link and email. Remove our name entirely on the Agency plan.
Bring your own keys
You pay OpenAI, Google and Perplexity directly at cost. Our fee is flat per client, so a heavy client never eats your margin.
Trend, not snapshots
30/60/90-day views that only draw a trend once there's enough data, and say "establishing baseline" until then.
Alerts that save the account
Brand vanishes for a cycle, competitor enters the top mentions, share drops past your threshold: you hear it before the client does.
QA you can show
Raw AI responses are stored. Spot-check any detection, flag a false positive, and the numbers correct themselves.
Built for the agency model
A new recurring line item, without new headcount.
Starter covers five clients for $99 a month, under $20 a client before your own API usage. Package AI visibility reporting at $150 a client and it pays for itself on the first one. Everything after that is margin, and the report carries your name.
- Costs scale with the clients you bill, not with query volume
- Reports auto-generate; an account manager reviews and sends in minutes
- Looks like a capability you built, because your clients only ever see your brand
| Semantic Scale, 5 clients | $99 |
| Your LLM API usage (typical, 5 clients) | ~$40–90 |
| Your cost | ~$140–190 |
| Billed to clients at $150 each | $750 |
API usage depends on your prompt count and repetitions; the default configuration runs 12 questions × 5 repetitions × 2 surfaces per client per week.
Honest by design
Measurement that survives a skeptical client. And their lawyer.
Selling “AI visibility scores” that aren't real measurements is a liability for you and your client. Semantic Scale is built so that every claim in a report is one you can stand behind.
Repeated sampling, always
Each question runs multiple times per surface per cycle because AI answers vary. One run is never a data point.
Sample size on every number
Percentages never appear without "X of Y answers" and a date range. Small samples are flagged as such.
Methodology on every report
A plain-language explanation of what was measured and a directional-signal disclaimer are built into every report and cannot be stripped by anyone.
Official APIs only
No scraping of consumer products, and no pooled credentials serving many agencies. Your keys, your account, your terms of service.
Pricing
Flat, per monitored client. Never per seat, never per query.
LLM usage is billed by each provider directly to you on your own keys. Switch plans anytime; monitoring never pauses on overage, you just get a nudge to upgrade.
Starter
Prove it with your first few clients.
- All four AI surfaces
- Weekly scans, 5× sampling
- Branded reports & share links
- Citation gaps & competitors
- "Powered by" footer
Growth
Most popularA service line across your book.
- Everything in Starter
- Alerts & thresholds
- Scheduled delivery with approval
- Detection QA workflow
- "Powered by" footer
Agency
Fully yours, at scale.
- Everything in Growth
- Remove all platform branding
- Raw data export (coming)
- Priority support & onboarding
Design-partner agencies get concierge setup and input on the roadmap. Ask about the pilot.
Questions agencies ask us
Does this guarantee my client will be recommended by AI?+
No, and anything that claims to is selling you exposure. Semantic Scale measures how often a brand appears across a defined sample of AI answers and how that changes over time. It's a directional signal that tells you where to focus, and every report says so.
Why do I need my own API keys?+
Two reasons. LLM providers' terms increasingly restrict one company's account serving many third-party businesses, so your own keys keep everyone compliant. And it keeps our fee flat: a client that needs heavy monitoring costs you API usage at cost, not a markup. We walk you through creating each key, one provider at a time, and validate it on the spot.
How close is this to what a real person sees in ChatGPT?+
Close, with caveats we state on every report. ChatGPT data comes from OpenAI's Responses API with web search, which OpenAI says runs the same model as ChatGPT search. Gemini uses official search grounding. Perplexity's API can pick different citations than a signed-in session. AI Overviews come from a licensed search-data provider because Google offers no API.
What does my client actually receive?+
A branded report (web link and PDF) with a plain-language summary, share of voice with sample sizes, trend over time, competitor comparison, the sources AI cited instead of them, and a methodology block. You can add your own commentary before it goes out.
Can I remove Semantic Scale's name completely?+
On the Agency plan, yes: no footer, no mention anywhere. Starter and Growth include a small "Powered by" line. The methodology and disclaimer block stays on every plan because it protects you.
Is my clients' data isolated?+
Every client lives under your agency with database-level row security. One client can never see another's data, and no other agency can see yours. When a client churns, you can delete everything about them in one action.
Be the agency with the answer.
Set up your agency, connect a key, add a client. Your first branded AI visibility report is a scan cycle away.