Start free, then pick a plan by its brands and monthly credits. You spend the credits observing your brands, on the assistants you choose and as often as your credits allow, and on AI agent runs at 80 credits on Claude or GPT, 40 on Gemini.
£0
For a first look at one brand.
Typical use: a £5 pack: 5 prompts on ChatGPT, Gemini and Perplexity, asked three times on each (six on Perplexity, at two settings).
Includes:
£149 a month
For one brand and one in-house marketer.
Typical use: a weekly default observation of one brand, on ChatGPT, Claude and Gemini.
Includes:
£399 a month
For up to 3 brands, or one brand observed more often.
Typical use: weekly default observations of 3 brands on ChatGPT, Claude and Gemini, or more often for fewer.
Everything in Starter, with:
£1,299 a month
For agencies and teams running up to 10 brands.
Typical use: weekly default observations of 10 brands on ChatGPT, Claude and Gemini.
Everything in Professional, with:
Custom
For single sign-on, dedicated support and procurement.
Everything in Business, on Enterprise terms:
How to choose: Free with a £5 pack, to see roughly where one brand stands; Starter, to track one brand over time; Professional, for up to 3 brands or one brand tracked more closely; Business, for agencies and teams running up to 10 brands; Enterprise, if you need security review, your own terms or invoicing.
Brand numbers are a ceiling, never a requirement: a one-brand team can take a higher plan to observe more often. Choosing a paid plan starts with a free account. You then confirm the plan and pay in Plans and billing; nothing is charged before that.
How far credits go depends on which assistants you pick, how many prompts you ask and how many audiences (simulated people) you add. Asking more prompts narrows the range on every result, and observing more often doesn't, so more prompts is usually the better use of credits. See what each observation size can detect.
Assistants cost us very different amounts to ask, so an answer from a dearer assistant uses more credits than one from a cheaper one. One credit is worth about 1p on a plan. You pay only for answers that arrive.
| Answer from | Credits for each answer |
|---|---|
| ChatGPT | 13 |
| Claude | 13 |
| Grok | 10 |
| Gemini | 9 (13 from 1 January 2027, when Google's price doubles) |
| Google AI Overviews | 6 |
| Google AI Mode | 6 |
| Perplexity | 1 |
One prompt on the default three assistants, asked three times each, uses 105 credits, about £1.05 on a plan. On all five assistants and Google's AI answers it uses 153, about £1.53. The default 27-prompt observation comes to about £28.
The Free plan has no monthly credits. Buy a pack when you want to observe. Subscribers can buy packs too, as a top-up.
| Pack | Credits | Roughly |
|---|---|---|
| £5 | 360 | 5 prompts |
| £10 | 740 | 10 prompts |
| £20 | 1,520 | 21 prompts |
| £50 | 3,850 | 53 prompts |
| £100 | 7,900 | Nearly 3 default observations |
First look: 5 prompts shows your rough position; it can't show change.
Every plan has every feature and unlimited people. Plans differ in brands, credits and depth: Free asks three assistants, without Google's AI answers. Enterprise adds terms, not features.
Scroll the table sideways to see every plan.
| Free | Starter | Professional | Business | Enterprise | |
|---|---|---|---|---|---|
| Price | £0 | £149 a month | £399 a month | £1,299 a month | Custom |
| Brands | 1 | 1 | Up to 3 | Up to 10 | Unlimited |
| Credits a month | None; buy packs | 14,900 | 39,900 | 129,900 | Custom, never below 1p a credit |
| Assistants | ChatGPT, Gemini and Perplexity | All 5 | All 5 | All 5 | All 5 |
| Google’s AI answers | Not asked on Free | Included | Included | Included | Included |
| Single sign-on and SCIM | Enterprise | Enterprise | Enterprise | Enterprise | Included |
| Dedicated support | Enterprise | Enterprise | Enterprise | Enterprise | Included |
| AI agents | From packs: 80 credits a run on Claude or GPT, 40 on Gemini | Same rates, from plan credits | Same rates, from plan credits | Same rates, from plan credits | Same rates, from plan credits |
On every plan: unlimited people; visibility, competitors and sources; every answer and source kept; audiences and product results; content fixes and publishing; email and Slack alerts; prompt demand, estimated from search data; before-and-after results for each fix; Readiness; AI crawler visits; impact from your analytics; client workspaces; reports under your own brand.
AI agents given buying tasks. Persipica records whether they find you, shortlist you and pick you, and the reasons they give. One run is one buying task given to one agent.
An estimate of how much a question is asked, from search data you connect: Search Console, Semrush or Ahrefs. It is estimated demand, not observed use of AI assistants.
Live checks on whether agents and AI crawlers can reach and read your site: product data, structured markup, feeds and robots rules.
Which AI bots actually visit your site, and which pages, from the server logs you import. Separate from Readiness, which checks what they could read.
Every feature, AI agents, client workspaces and reports under your own brand among them, and unlimited people, on every plan. Plans differ in brands and monthly credits. Paid plans choose from five assistants with their own web search (ChatGPT, Claude, Gemini, Perplexity and Grok), plus Google's AI answers (AI Overviews and AI Mode), which count separately; observations start with ChatGPT, Claude and Gemini, and you can add the rest. The Free plan observes in less depth, on ChatGPT, Gemini and Perplexity without Google's AI answers. Enterprise adds terms rather than features: single sign-on and SCIM, and dedicated support.
Each prompt is asked three times on every chat assistant (six on Perplexity, at two settings), and every result comes with its range, its sample size and the answers behind it. That costs more to run than asking once, and we price on what the answers cost.
Failed, timed-out and blank answers are never charged. This is automatic; you don't need to ask.
You pay only for the answers already received, and you get partial findings that show exactly those answers. Prompts that hadn't been asked yet cost nothing. If we stop an observation, it costs you nothing.
Tell us within 30 days and we return the credits to your current month's allowance.
You can cancel any time, and your plan stops at the end of the month you've paid for. Fees for a month already paid aren't refunded, except where the law requires it.
No. No VAT is added. The price shown is the price you pay.
Yes. AI agents are on every plan, and a run is charged in credits when it ends: 80 credits on Claude or GPT, 40 on Gemini. A run that fails is free. Runs use the same monthly credits and packs as observations, and the app shows what a run costs before it starts. One run is one buying task given to one agent. A task run 120 times, as in the home page's example, would use up to 8,000 credits if split 40 each across Claude, GPT and Gemini, about £80 on a plan. Run prices are set from modelled costs and may change once runs have been measured.
Yes. Start on the Free plan with no card, and buy a £5 pack for a first look. You can also check one question for free without an account.
Yes. Move up or down at any time from Plans and billing in the app.
Send your brand and three competitors. We observe your brand and walk you through the findings in 30 minutes, including what the ranges mean. It's free, and we reply within one business day.