Seoptist vs doing it in-house with spreadsheets

The in-house version of AI visibility tracking usually looks like this: a sheet with one row per prompt, columns for each engine, a cell for "mentioned yes/no" and maybe a notes column for who else came up. Someone fills it in on a Friday afternoon. It is honest, cheap and better than nothing, and plenty of businesses have learned something real from it.

Where spreadsheets work

For a one-off check, a spreadsheet is the right tool. If you want to know today whether five assistants mention you for twenty prompts, an hour with a sheet will tell you, and you will read every answer in full while you do it, which is itself educational.

Spreadsheets also suit a small pilot before committing budget. Running the manual version for a month shows whether the numbers move at all and whether anyone in the business looks at them. If nobody does, no tool will change that.

A developer calling the official APIs into a sheet can get surprisingly far, too.

Where it breaks down

The hidden cost is staff time, and it compounds. Twenty prompts across three engines with three samples each is 180 answers a week. Reading them, deciding whether a mention counts, noting position and tone, extracting the cited domains and checking whether the price quoted matches your price list is several hours of careful work. It gets skipped the first busy week and never quite restarts.

Consistency is the second problem. Two people classify "mentioned" differently. Someone uses their personal ChatGPT account with memory on; someone else uses the API. Last month's sheet used different prompts. Trends built on that are not trends.

Then there is everything a sheet does not do: it does not know that you blocked GPTBot in robots.txt last March, that your LocalBusiness schema has the old phone number, that your service pages render only in JavaScript, or that your INP on mobile is poor. It does not produce a report a client can open, and it does not alert anyone when an assistant starts quoting the wrong opening hours.

Side by side

Criteria	In-house with spreadsheets	Seoptist
Engines covered	Whatever someone has time to check, often one or two	ChatGPT, Gemini, Perplexity, Claude, Google AI Overviews, AI Mode, Copilot (3 on Starter, 5 on Growth, all 7 on Agency and Enterprise)
Sampling	Usually one answer per prompt, when the sheet gets filled in	3 answers per prompt per engine every week, automatically (5 answers and daily runs on Enterprise; daily add-on for 20 prompts at £49/month)
Accuracy alerts	Manual comparison, if remembered	Answers compared against your fact card (prices, address, phone, product details) with alerts on mismatch, Growth and above
Audit	Separate tools, separate sheets	Technical SEO, Core Web Vitals and GEO readiness audit (100 pages Starter, 500 Growth and Agency, 5,000 Enterprise)
Checklist	A to-do tab someone maintains	One prioritised list scored impact x 2 minus effort, with how-to-fix guidance and copyable code, auto-verified on the next crawl
White-label	A logo pasted into a chart export	Agency plan (£399/month) and Enterprise: logo and colours, client workspaces, client-viewer logins
Price from	Free software, plus your staff time and any API spend	£49/month ex VAT, 14-day trial with card, two months free on annual billing

If you do build your own, use the official APIs rather than scraping chat interfaces; the latter breaks accounts and terms of service. Seoptist collects answers only through official APIs (OpenAI, Gemini, Perplexity Sonar, Anthropic) and licensed data providers (SerpApi, DataForSEO) for the same reason.

Who should choose which

Stay with the spreadsheet if you need a single snapshot, are testing whether the topic matters to your business, or have a developer with spare time and a genuine interest in building the pipeline. None of those are bad reasons.

Replace it when the Friday-afternoon routine has lapsed twice, when you need more than one engine on a schedule, when you want the technical audit and the AI numbers to feed one list of things to fix, or when the output has to go to a client or a manager who expects a report rather than a tab. The honest comparison is not £49 against free; it is £49 against the hours the sheet quietly consumes and the weeks it quietly misses.

The free AI Visibility Check at /free-check gives you a snapshot without building anything. If you then want it to run every week without you, plans are at /pricing.