Mock data for AI-driven enterprises

Give your leaders
a company to practise on.

Every executive wants their people leaders using AI against the company’s data. Almost none of them want that on the first attempt, in production, with the real customer list. So build the same company again, with none of the risk, and hand it over.

No account needed to look. You only sign in when there is a database to build.

1

domain typed in

40+

vendor systems recognised, from Salesforce to SAP

0

real customer records involved

24h

free sandbox, then the database is spun down

What happens

Four steps, one URL at the end

01

We find the company

Brandfetch gives us the real logos and colour palette. Use the arrow keys to pick the one that matches, hit Enter, and the whole flow is branded like the company from that point on.

02

We work out what they run

BuiltWith, the fingerprints in their own page source, vendor names in their press releases and job postings, and a language model to fill the gaps in the back office. Every system comes with the evidence behind it. Where we cannot tell, we ask.

03

We build the world

A private Postgres database, one set of tables per system with the vendor's real object and field names, and records generated against what the company actually sells, to a customer base that is entirely fictional and internally consistent across every system.

04

You get the URLs

One MCP server per system, plus a knowledge graph built from the company's public web presence. Paste the URL into Claude, ChatGPT, Cursor, or anything else that speaks MCP. There is a REST API and generated OpenAPI documentation too.

Why this exists

The risk is not that your people will use AI badly on real data. The risk is that they never get enough repetitions to use it well at all.

A leader learns what an agent can do by pointing it at a pipeline and asking an awkward question. That takes a real CRM with real shape: stages that mean something, deals that slip, an account that is quietly escalating three critical cases while its renewal stalls.

Generic seed data does not teach that, because nothing in it is connected to anything else. Here the same fictional customer appears as a Salesforce account, a NetSuite invoice, a Zendesk organization and a complaint on the company’s Facebook page. The arithmetic adds up. The funnels narrow. Closed deals have a reason.

What is real and what is not

Real: the company’s brand, its products, its business lines, its public leadership, the systems it runs, and everything in the knowledge graph, which comes from pages it published itself.

Not real: every customer, every contact, every transaction, every ticket, every number. Customer names are deliberately fictional so nobody mistakes this for a data leak.

Never touched: the company’s actual systems. Nothing here authenticates against any real vendor. There is no path from this product to production.

Plans

Start free. Pay when it needs to persist.

Sandbox

Free

One system, 24 hours, no card.

  • One mock system of your choice
  • MCP server plus REST API for that system
  • Company knowledge graph, read only
  • Data explorer and generated schema docs
Details
Persistent

$49.99 / month

Keep the data. Five systems.

  • Up to five mock systems per instance
  • Data stays live as long as the subscription is current
  • Three concurrent instances
  • Knowledge graph refresh on demand
Details
Enterprise

From $999 / month

Many people, one shared world. Priced per engagement.

  • Everything in Live
  • Multiple named users on the same instances
  • Shared MCP endpoints across a cohort or a leadership team
  • Org-level roles, audit log and SSO on request
Details

The Persistent plan comes with a one month free trial. Enterprise pricing starts at $999 a month and is quoted per engagement.