Public AI answers

Answers for the public, drawn from your own published content and carrying a link to the source. We build this onto your site and your material — the shape it takes depends on what your users ask and what you publish.

What it is

The usual starting point is a pop-out chat panel: streaming responses, markdown, sessions that persist across pages. That is one shape. Some clients want answers inline on the search results page, some want them on a single service page and nowhere else, some want a different escalation route when the answer is not good enough. The retrieval and the grounding are the same underneath; the surface is built to fit.

The demo is an example

What we show in a demo is one build over one corpus, so you can see the citations and judge the answer quality. It is not a product you are buying a licence to. Yours would be built against your content, your tone, and the questions your users actually arrive with.

What people ask it

“When is my next bin day?” “How do I object to a planning application?” “What housing support is available?”

And return immediate, citation-backed responses — using your own content.

In production

Multilingual health information. Translated medical resources in 60+ languages. A user asks for COVID vaccine information in Arabic and gets the right document in the right language.

How we start

PretaGov offers:

• Rapid AI search pilots using your existing content
• Low-risk, fixed-cost setup

Looking for search across internal material instead — your own documents, dashboards and records, never leaving your infrastructure? That is Onyx enterprise search.

You can try this on three corpora in the live demos. It is one of four things we do under AI integration.