Persipica

Everything Persipica does

Persipica measures how AI agents and AI assistants treat your brand and products, turns what it finds into changes you can make, and checks them on the next observation.

The views below use a sample brand. Every number in Persipica opens the answers behind it.

AI agents

Agents get real buying tasks. Persipica records whether they find your brand, shortlist it and pick it, compared with each competitor, and why they chose someone else.

  • Found, shortlisted and picked, next to each competitor
  • The reasons agents gave for their choice
  • Agent runs use credits, and the app shows what a run costs before it starts

One run, step by step

Illustration with made-up brands

Find merino walking socks that stop blisters, under £20 a pair

2. Read

It opened each product page and size guide:

  • Ridgeway Trail Merino, £16.00: sizes in text, in stock
  • Fellside Peak Merino, £18.00: size guide in an image it couldn’t read
  • Tor Outdoor Hiker, £14.00: 40% merino
  • Pennine Mills Junior, £10.00: children’s sizes

3. Shortlist

It shortlisted Ridgeway, Fellside and Tor Outdoor, and dropped Pennine Mills Junior, a children’s sock.

4. Pick

It picked Ridgeway Trail Merino, £16.00, and said why:

“Ridgeway lists its sizes in text and shows them in stock. Fellside’s size guide is an image I couldn’t read.”

Averaged across 40 runs each on Claude, GPT and Gemini, Fellside was picked in 22% (likely 12–38%), Ridgeway in 38% (likely 24–54%): not a clear difference. 3 of the 120 runs broke off before a final pick.

Visibility

How often answers name and cite you, by assistant and against your competitors, week by week.

  • Mention rate, citations, share of voice and position
  • Each rate with its likely range and the number of prompts it rests on
  • Hover a week to read what the assistants actually said

Fellside, named in AI answers

Last 8 weeks
Fellside, with its likely rangeRidgeway, top competitor

Across all four assistants, mentions went from 26% on 17 Aug to 51% on 5 Oct. The ranges still overlap, so it’s not a clear difference yet.

Audiences

Describe the people you sell to. Persipica sends prompts written in each one's voice, with follow-ups, and shows who hears about you and who hears a competitor.

  • Audiences add prompts to the same observation
  • Every attribute labelled as your data, a public signal or an assumption
  • Simulated people are always labelled as simulated

Who hears about Fellside

ClickTap an audience for detail

Ridgeway is named at 44%, likely 17–74%. Asked as: "Which walking socks stop blisters on long days?"

Tor Outdoor is the main alternative at 38%, likely 14–70%. Cheaper multipacks are mentioned in most answers.

Pennine Mills is named at 79%, likely 45–95%. Answers cite a review site's "longest-lasting kids' socks" list, which doesn't include Fellside.

Fellside, all four assistantsTop competitor

Products and prices

Bring in your catalogue from Shopify, a Google Merchant Center feed or a CSV. See which products AI names, the prices it quotes against yours, and the stores it sends shoppers to.

  • Results for each product, not just the brand
  • Wrong prices flagged with the page they came from
  • Works without a catalogue too, from your brand and website

Your products in AI answers

Share of the answers that name Fellside, averaged across all four assistants. Prices checked against your catalogue.

£18.00 quoted correctly in 58 of 61 answers. Shoppers are sent to your store and to two outdoor retailers.

£9.00 quoted in 2 of 6 answers, your price is £11.00. Both are Gemini's, reading last season's price from two retailer pages.

£24.00 quoted correctly in 20 of 20 answers. Most often recommended to students, usually next to Tor Outdoor's multipack.

Accuracy

Persipica flags answers that quote a price or stock level your catalogue contradicts, or cite a source that doesn't say what the answer claims.

  • The wrong claim highlighted in the answer itself
  • The source it most likely came from
  • Alerts when answers start getting facts wrong

Answers that get you wrong

From this week's observation

"For kids, Fellside's Trail Junior costs about £91 and comes in three colours, but it's sold out in most sizes2, so try Ridgeway."

  1. Price. Your price is £11.00. The £9 comes from a retailer's old page.
  2. Stock. Your catalogue shows Trail Junior in stock in every size.

Fixes

Findings become a list of changes your team can work through. Persipica drafts each one from the answers where you were missed, and later observations show the before and after.

  • Page rewrites, FAQs and comparison pages, drafted for you
  • Publish to WordPress, Webflow, Contentful or Sanity, or anywhere with a webhook
  • Owners and status for each change, plus your own
  • Before and after for each change, with ranges. A change is called one only when the ranges don't overlap

What to change next

Before and after, with ranges

Your robots.txt turned Perplexity's crawler away, so it couldn't read your pages.

Observed. Perplexity named Fellside in 18 of 144 answers on 7 Sep (13%, likely 4–31%) and 82 of 144 on 14 Sep (57%, likely 38–74%), across its 24 prompts. The ranges don’t overlap, so it’s a clear difference, but not proof the change caused it.

Observed. Averaged across 40 runs each on Claude, GPT and Gemini, agents shortlisted Fellside in 32% of runs before the change (likely 19–47%) and 41% after (likely 27–56%).

What it means. The ranges overlap, so it’s not a clear difference. Even a clear change wouldn't show that this fix caused it.

Two retailer pages still show £9.00. Persipica checks again on the next observation.

Also in Persipica

The tools around each observation.

Look deeper

  • Readiness

    Whether agents and AI crawlers can read your site.

  • AI crawler visits

    Which AI bots visit, from your server logs.

  • Prompt demand

    Prompts ranked by estimated search demand, not AI usage.

  • Competitors and sources

    Who is named instead, and the sites answers cite.

Share the results

  • Answers

    Every answer kept, with its sources.

  • Reports

    Findings, an executive view and share links.

  • Impact

    Visits and revenue from AI, from your analytics.

Run it your way

  • Alerts and schedules

    Observations on a schedule you set; alerts show the raw change between two observations.

  • Teams and agencies

    Many brands, client workspaces, roles, single sign-on and a security log.

  • API and MCP server

    Pull results into your tools, or ask Persipica from an AI assistant.

See it on your own brand

Start free, or see what each plan includes.

Start free See pricing