The Prompting Company vs Promptwatch: how each would measure your visibility score
Our visibility score is the share of agreed buyer prompts where AI engines name you. How The Prompting Company and Promptwatch would each measure it, given their prompt caps, engine gating, and answer budgets.
Everything we report comes down to one monthly number: the share of an agreed set of buyer prompts in which AI engines name you. A number like that is only as good as the tool collecting it. So when a new platform turns up in a buyer's shortlist, we ask a narrow question. Could it produce the score, and how trustworthy would the result be?
The Prompting Company is the latest one we have looked at this way. It is aimed at developer companies and does plenty beyond visibility tracking, but here we only look at measurement.
Sources: promptingcompany.com pricing, home, and llms.txt pages plus the co-founder's published method, checked on 24 October 2026; Promptwatch's llms.txt, re-checked on 21 October 2026.
Its share of voice is our score
The Prompting Company's co-founder describes the method like this: run about 100 long-tail buyer prompts across four engines (ChatGPT, Perplexity, Google AI, Claude) daily, and define share of voice as the percentage of runs that mention the brand.
That is the same fraction we call the visibility score. It is not what we call share of voice. In our reports, share of voice is your mentions as a share of all mentions of the competitors you track, so it moves when rivals gain or lose. A percentage of runs naming you moves only when you do.
This matters when you read vendor case studies. The Prompting Company says Hyperbrowser's share of voice in browser-automation AI answers rose from 3% to 58%. That is vendor-reported, and we have not checked it, but under the co-founder's definition it describes a visibility score, not a competitive split. Translate the term before you compare it with anything in your own reports.
Prompt count sets how sharp the score is
If you track 25 prompts, each prompt is worth 4 points of score. Lose one and the monthly number drops 4 points, even if nothing meaningful changed. With more prompts, a single swing matters less.
| Plan | Price | Prompts | Points per prompt | Monthly answer budget | Budget per prompt |
|---|---|---|---|---|---|
| TPC Basic | $99/mo | 25 | 4 | 1,000 credits | 40 |
| TPC Pro | $299/mo | 100 | 1 | 4,000 credits | 40 |
| Promptwatch Essential | $95/mo | 50 | 2 | 6,000 responses | 120 |
| Promptwatch Professional | $245/mo | 150 | about 0.7 | 18,000 responses | 120 |
| Promptwatch Business | $579/mo | 350 | about 0.3 | 42,000 responses | 120 |
The last two columns are our arithmetic from the published limits. The Prompting Company charges one credit per AI response per region, so a client tracked in two markets spends credits twice as fast.
Repeat runs keep the score stable
AI answers vary between runs, which is why we never score a prompt on a single check. The budget per prompt decides how many checks you get. Promptwatch carries three times the budget per prompt, though it spreads that over far more engines. The sharper point is that on The Prompting Company's lower plans, credits run out before prompts do.
The co-founder's own method shows the pressure. Run 100 prompts on four engines every day and, if each engine's answer costs one credit, that is roughly 12,000 credits a month. That is the White-glove allowance ($3,000/mo), not the 4,000 on Pro, the plan that actually carries 100 prompts.
Engine gating changes what "named by AI" means
A visibility score has to say which engines it covers, and it should be broken out per engine, because buyers do not all use the same assistant.
On The Prompting Company's comparison table, Basic covers ChatGPT, Perplexity, and Gemini. Pro adds Google AI. Claude and DeepSeek appear from White-glove, with coding agents (Claude Code, Codex, OpenCode). The Basic plan card says "Access to all models", so the page contradicts itself; we go by the table. Claude is also part of the co-founder's four-engine method but sits above Pro in the table.
Promptwatch runs every tracked model on every paid plan: ChatGPT, Gemini, Claude, Perplexity, Grok, Llama, DeepSeek, Mistral, Copilot, Google AI Overviews, and AI Mode. A score built on Essential and a score built on Business cover the same engines, so moving plans does not move the number.
What sits outside the score
The Prompting Company's distinctive output is an agent experience score: a coding agent tries to use your product from its docs in a sandbox, and a judge grades the result. It is a useful number for an API company. It is not a visibility score, and we would keep it out of the headline rather than blend the two. The audit behind it is Enterprise only.
Where we measure it
We measure the score in Promptwatch. Prompt tracking gives a per-engine score across every model, the stored responses let us open any prompt and read exactly how you were named, and citation analytics shows which pages moved it. Share of voice sits underneath, against a competitor set agreed at the start. If you are a devtools company and also want to know whether coding agents can use your product, The Prompting Company's agent experience work is worth a separate conversation. It answers a different question.