Logo

Scope – Agent Customer Journey Analytics for AI Agents

Scope is an agent experience platform that shows which AI agents find your product, what they came to do and where they stop. Y Combinator-backed; no public price list, book a demo at tryscope.com.

Scope – Agent Customer Journey Analytics for AI Agents

Scope – Agent Customer Journey Analytics for AI Agents

Scope is an agent experience platform that shows which AI agents find your product, what they came to do and where they stop. Y Combinator-backed; no public price list, book a demo at tryscope.com.

Features

Pricing

  • Pricing

Description

Scope is an agent experience platform that claims to show which AI agents find your product, what they came to do and where they stop, then rank the fixes by the revenue those drop-offs put at risk. Official pages describe a single journey per agent across your docs, API and MCP server — from the search that brought it in, through install, sign-in and the first API call or payment. The company behind the site uses the trading name Scope, is based in San Francisco and is backed by Y Combinator (Spring 2026 batch, according to its Y Combinator company profile, which also lists a 2025 founding year). This is not a prompt-citation or AI-overview tracker in the usual GEO sense, and it is not a replacement for analytics built for human page views.

There is no public price list. A /pricing page on the official site returned 404 when checked on 13 November 2026 (Europe/London), and no free plan or self-serve trial was advertised. The official door is book a demo: you give a work email, pick a time, and the vendor says it will run Scope against your product live. Product screens on the homepage are labelled as a sample workspace, so the session counts and revenue figures shown there should be read as illustrations, not customer results. Full write-up: Scope review 2026.

Who It’s For

  • Developer-tools, API, infrastructure, CLI and MCP teams whose products coding agents can already sign up for, install or call — the official FAQ names developer tools, APIs, payments, SaaS, commerce and travel as the intended audience.
  • Growth, product, DevRel and documentation teams who suspect agents are reading docs and hitting APIs, but whose existing analytics never record those visits because agents do not run page scripts.
  • Teams that can ship fixes to docs, auth, examples or MCP descriptions and want those changes ranked by sessions lost and tested on real agents before buyers meet them.
  • Skip or use with care if you need a published rate card or a self-serve trial, you only want a dashboard of how often ChatGPT or Perplexity mentions your brand, or you expected a drop-in replacement for Google Analytics, Amplitude or PostHog on human traffic.

Core Capabilities

Attribute the agent, and how it found you

Official copy says Scope names arriving agents — Claude Code, Codex, Cursor, Muse, Instinct and a long tail — and shows what the agent searched for, the page it landed on and the ones it skipped. Every name carries a confidence label: Verified (signed by the agent’s provider), Declared (the agent said who it is) or Inferred (its behaviour matches a family). Anything the vendor cannot name is grouped by how it behaves, rather than guessed. Sessions are also split by where the agent runs.

Read the task, then group it into intents

The vendor claims every session carries the task the agent was given, in the person’s own words. Scope groups those tasks into intents and says it can show which intents turn into customers. That is a stronger claim than a simple user-agent flag in a log, and it should be tested in a demo on your own product.

Join the steps into one journey

Docs read, package installed, signed in, first API call, MCP connected: official pages describe Scope joining each step into one journey and marking the moment it broke. A sample trace on the homepage walks a Codex session from reading /llms.txt and a billing doc, through npm install and login, to a repeated 401 on an API key header, then a stop with no first payment. Treat that walkthrough as a product illustration.

Tie journeys to conversion and revenue

The vendor says you can see conversion and revenue by agent, by intent and by page, and an estimated monthly figure for what each unresolved drop-off costs, based on the conversion rate of sessions that got past it. How revenue is joined to a session (billing system, product events, or something else) is not spelled out on the public site, so ask in the demo.

Rank fixes, then test them on real agents

Scope claims to rank fixes by the revenue they unlock and to show each one as a readable change — a corrected code sample, a docs page or an MCP description — ready to review and ship. It then says it tests the change on real agents before your buyers meet it. Official research posts also argue that agents often skip a vendor or fail during setup; those figures are the vendor’s own tests, not an independent benchmark.

Key Features

  • Agent identification with Verified, Declared or Inferred labels, plus a grouped unnamed tail.
  • Session reconstruction across docs, APIs and MCP servers, including install, sign-in and first payment.
  • Intent clustering from the task text the agent was given.
  • Conversion and revenue views by agent, intent and page (vendor claim; joining method unpublished).
  • Drop-offs ranked by estimated revenue at risk, with suggested docs or API fixes.
  • Pre-release tests of those fixes against real agents (vendor claim).
  • A public database agent-readiness leaderboard and a short research blog on agent experience (AX).
  • Demo booking as the only public way in; no self-serve checkout found.

Plans & Pricing

Checked 13 November 2026 (Europe/London) on tryscope.com. No plan names, seat prices or usage tiers were published.

Plan Price What official pages say
Book a demo No public price The homepage and /demo ask for a work email, then a time. The vendor says it will run Scope against your product live: which agents reach it, where they stall, and what each drop-off costs.
Self-serve plans or a trial Not advertised No /pricing page (404). No free plan, credit pack or trial CTA was found on the official site, sitemap or llms.txt index.

Pros & Cons

Pros

  • Clear, timely brief: measure the path an agent takes on your product, not only whether a model mentioned you.
  • The five-step loop — attribute, intent, journey, revenue, improve — is easy to explain to a product or DevRel team.
  • Confidence labels on agent names (verified, declared, inferred) are more careful than a single guessed user agent.
  • Public research and a database leaderboard give a sense of how the vendor thinks about agent experience.
  • Backed by Y Combinator, with a focused early audience in APIs, infrastructure, CLIs and MCP servers.

Cons

  • No public rate card, trial or self-serve checkout, so you cannot compare cost before a sales conversation.
  • No privacy policy or terms page was found on the official site (both URLs returned 404), which is a gap for a product that reconstructs tasks and ties sessions to revenue.
  • Homepage metrics are labelled sample data, and no customer case studies were published on the pages we checked.
  • Research figures (agents skipping a vendor about one time in three, and a 28% failure rate even when the product is named) are the vendor’s own tests.
  • Early-stage company (founded 2025; Y Combinator Spring 2026), so the product and commercial terms may still move.
  • The name Scope is generic, and the product is easy to confuse with unrelated “scope” tools in search.

Where It Fits

Scope sits with AI SEO and GEO tools because it is about how agents discover and choose a product, and with analytics because the live pitch is a journey and revenue console, not a prompt-tracking grid. For dashboards that measure how often assistants mention or cite a brand, see Nimt.ai, ChatRank, OptimizeGEO and Peec AI. For a free page-level SEO and AEO audit, including whether AI crawlers can read a site, see Agentic SEO Check. For a Google-focused SEO agent, see ChatSEO. For a managed AI-visibility retainer, see FirstIndex. Comparing options? Read our 25 best AI SEO tools for 2026 guide, or browse the directory of AI tools for digital marketers.

FAQs

What does Scope do?

Official pages call it the agent experience platform. It aims to show how AI agents discover, evaluate and use your product, join each session into one journey, tie that journey to sign-ups and revenue, and tell you what to fix so more of them succeed.

How much does Scope cost?

We could not find a public price. The official call to action is to book a demo at tryscope.com/demo. Treat any third-party guess at a monthly fee as unverified.

Is there a free trial?

Not on the public site we checked. There is no trial CTA, free plan or self-serve signup in the header, sitemap or demo page.

Is Scope a GEO or AI-search tracker?

Only in part. It sits next to GEO tools because agents often start with a search, and the company has written about AI-search discovery. The product pages we reviewed centre on agent sessions across docs, APIs and MCP — install, sign-in, first call and payment — rather than a share-of-voice dashboard for ChatGPT or Google AI Overviews.

Which agents does Scope identify?

The official FAQ lists Claude Code, Codex, Cursor, Muse, Instinct and others, with new agents added as they appear. Names are labelled verified, declared or inferred, and unnamed traffic is grouped by behaviour.