Tonecast for agencies and analysts
For an agency, Tonecast is a way to tell each client what AI assistants say about them, why, and what to do about it, with numbers the agency can defend. The Agency plan gives one workspace per client under one company account, and every figure opens down to the answers and web pages behind it.
Tonecast is a SaaS platform that shows what AI assistants (ChatGPT, Claude, Gemini and Perplexity) say about a brand, product, person or topic, and which web pages shaped those answers. This page is for two people in an agency: the account manager, who reports to clients and needs each client's data kept apart, and the analyst, who refines prompts, segments the results and exports raw data for their own analysis.
Which questions does Tonecast answer for an agency?
| Question | Where the answer is |
|---|---|
| How is each client doing in AI answers this month? | Home: one card per campaign in the client's workspace, with visibility, tone and early-signal state |
| Which engine, language or country is the problem? | Campaign Overview and AI Answers, filtered by engine, language, country and prompt |
| Which pages should the client work on? | What shaped it (source graph) and Actions (gap & action list) |
| Which facts do AI assistants get wrong about the client? | Claims view, against the client's fact sheet |
| How were these numbers produced? | Methodology card: engines, samples, intervals, measured accuracy |
| Can I analyse the raw data myself? | CSV exports of answers and source graph, with the current filters |
How are clients kept apart?
On the Agency plan, one company account holds up to 10 workspaces, typically one per client. Each workspace has its own members, so a client's team can be invited to their workspace only. Campaigns, prompts, fact sheets and results never mix between workspaces.
What does the analyst get?
- Prompts by category (discovery, comparison, evaluation, price, support, current events), editable at any time, tracked per language.
- Filters by date range, engine, language and country on the campaign overview; by prompt, engine and tone on the answers; and by tier, source type, engine, tone, minimum weight or domain on the source graph. Filters change what you read, never what is collected.
- Drill-down from every number to the answers and the source pages that compose it.
- Corrections: mark a page as not about the subject, or correct a tone. Corrections apply to the indicators immediately, are kept apart from the model's scores, and improve future models.
- CSV exports of answers and source graph with the filters applied.
Can an agency track topics and public figures, not just brands?
Yes. A campaign can be about a brand, a product, a public person or a general topic. For a topic, Tonecast proposes questions such as "what is happening with…" and "what are the positions on…", and the source graph shows which pages frame the debate in each language.
How do I report to clients?
Today, the campaign views and CSV exports cover reporting, and the methodology card answers the question every client eventually asks: "how do you know?". White-label reports, shared views and a public API are part of the Agency plan and are being rolled out; the features page tracks what is live.
Which plan fits an agency?
The Agency plan (€549 per month, or €5,270 per year) includes 10 workspaces, 30 active campaigns, 200 tracked prompts, up to 5 engines and 10,000 source pages per workspace. Smaller agencies with a few clients can start on Pro (€249 per month, 10 campaigns, one workspace). Prompt packs add 25 prompts for €25 per month. See pricing.