What a Cambridge Spinout Should Build First for Search and AI

A spinout out of a Cambridge Enterprise licensing process usually arrives with a strong reputation and no public record of its own. The science is documented under the university's name. The company itself has a placeholder domain, a LinkedIn page, and very little a crawler or an assistant can read.

That position is better than it sounds. No legacy content needs correcting and no historic description needs unpicking. The first properly written page is therefore the definitive one. The risk is leaving the placeholder in place for a year while the fundraising runs.

Timing matters more than polish at this stage. A domain that has carried four accurate pages for six months is easier for an engine to trust than a beautiful site published the week before a funding announcement. Publish early, then improve the pages as the company learns what its buyers ask.

1.Cambridge Enterprise is the University of Cambridge's commercialisation arm, handling licensing, spinout formation, and early investment for university-derived technology.

1. A homepage with one plain sentence

   The first sentence states what the company makes or does, in the words a buyer or an investor uses. The technical detail follows immediately underneath. This sentence is what both a search engine and an assistant lift when they summarise you.

2. A technology page written for two readers

   An expert needs the mechanism, the data, and the boundaries of the claim. A generalist needs to know what problem it solves. Write the second at the top and the first below it, on the same page.

3. A team page with real names and histories

   Founder names, prior institutions, and roles give a model something to corroborate against public records. It is also the page investors read first. Publishing it accurately costs nothing.

4. A contact page with the full registered address

   Registered office, company number, and a working email. This page resolves the identity questions everything else depends on.

State it factually and once. Name the technology licensed, the department or institute it originated in, and the year the company was formed. A short, accurate paragraph does more good than a page of association language.

Vagueness here is expensive. A company that describes itself only as Cambridge-affiliated invites the exact confusion that costs it citations later. A model reading that phrase cannot tell whether the entity is the university, a department, or an independent business. Precision is the cheapest disambiguation available.

Respect the boundaries as well. Publish what the licensing agreement and any confidentiality terms permit. Check anything uncertain with the technology transfer office before it goes live.

Organization schema is the priority, carrying the legal name, the trading name, the company number, the registered address, and the profiles you control. Those fields are what an engine cross-checks against Companies House and LinkedIn. Agreement between the three is the point of the exercise.

Keep the robots file permissive from day one. GPTBot, PerplexityBot, ClaudeBot, and Google-Extended all need access. A default template that blocks them is a silent and completely avoidable loss.

Submit a sitemap at launch and check indexation coverage a fortnight afterwards. A new domain is exactly when an indexation fault is cheapest to find.

Defer the blog. A publishing schedule nobody has capacity for is worse than no blog at all. An abandoned news section dated eighteen months ago reads as a dormant company.

Defer case studies and testimonials until there is real permission and a real result to describe. Inventing either is the fastest way to lose a technical buyer.

Defer gated content entirely at this stage. Material behind a form is invisible to every engine. A spinout with almost no public record cannot afford to hide the little it has.

Defer the rebrand as well. A name change after the entity has begun to resolve costs you every piece of corroboration you have accumulated. The naming decision belongs at formation.

It does not claim that four pages produce enquiries. They produce a resolvable entity. That is the precondition for everything else rather than a result in itself.

This page sits inside the Cambridge deep-tech field guide. The mechanism behind AI citations is covered in how AI assistants decide which Cambridge firm to cite. The diagnosis order for an existing site is on the zero-presence page.