See how AI could work inside your business — using your own approved information

A structured evaluation experience, not a free chatbot. We may prepare an initial controlled demonstration from appropriate public information on your organisation's website and other suitable public or supplied business sources. Over thirty days, assumptions are identified, knowledge is verified and refined with you, your nominated team uses the assistant privately, and you finish with a written record of what was tested and what was not. No fee, no obligation, no payment details.

Start your free 30-day experience

Or request a consultation first

Your information makes the test meaningful

Most AI conversations stall in the same place. Everyone can agree in principle that it might help. Nobody can say what it would actually do here, with our information, for our people, answering our questions.

This is our answer to that. We agree one narrow problem and a representative set of questions before reviewing any material. For example, the assistant might help nominated staff find answers in one approved procedure set; it would not become a general staff adviser. You get a controlled demonstration — narrow, bounded, and built from appropriate public or supplied business information that is verified and refined with you — so the decision about whether to go further can be based on observed use.

A small way to test a real problem

Choose one problemTogether we define one narrow problem, the intended users, a representative set of questions and the human escalation point. You then decide which limited, approved information your organisation will provide for that scope.
Review the starting informationWe review appropriate public or supplied business information and separate what is verified from what is assumed or needs checking. Material assumptions come back to you before they are treated as authoritative knowledge.
We build and testWe build the assistant, set the rules for what it must say and must refuse, then test it against a written protocol. Every result is recorded, pass or fail.
Your nominated people use itWe show you the assistant working, including the questions it declines, and we hand over what the testing found.
Review what you learnedYou and the people you nominate use it privately for thirty days. At the end we review what you learned and you decide what, if anything, happens next.

Note on your time

Realistically: one conversation to scope it, one to be shown it, and some of your time in between confirming that information is correct. The verification step is the part that needs you, and it is the part that determines whether the result is any good.

Included and not included

Included

One private, controlled demonstration assistant for one agreed narrow problem, built from appropriate public or supplied business information and refined through your review
Review of that information, with anything uncertain flagged back to you
A representative set of the questions your people or your enquirers actually ask
Written rules for what the assistant must say and must refuse
Testing against a written protocol, with results recorded
One presentation walking you through it
One reasonable round of corrections
Thirty days of private access for your nominated team
A written record of what was tested and what was not

Not included

Deployment on your website or any public channel
Integration with your systems, databases or records
Handling of personal, payment or health information
Compliance, regulatory or certification assurance
Unlimited revisions, or support after the thirty days
Any guarantee of a business result
A finished production system
Ongoing hosting or maintenance
A free chatbot, or any always-on public service

Three rules we will not compromise on

Nothing is placed on your website, or shown to your customers, residents, members or the public, unless you approve it separately and in writing. The trial does not touch anything public.
The assistant runs in a restricted private environment, available to the people you nominate. It is not linked from anywhere and not indexed.
No unnecessary personal information, highly sensitive health information, payment-card credentials, passwords, authentication secrets or private API keys go into the demonstration. A different use case requiring personal or sensitive information needs separate assessment and appropriate safeguards.

An initial demonstration may use appropriate public information from your organisation's website and other suitable public or supplied business sources. Public availability does not remove privacy or intellectual-property obligations. We identify reasonable best-fit assumptions where relevant, and demonstration content must be verified and refined through discovery before it is treated as authoritative production knowledge. Coverage remains deliberately narrow — and when a question falls outside it, the assistant says so instead of guessing.

Read our approach to responsible AI

What the 30 days are designed to tell you

This is a structured evaluation experience and a controlled demonstration. It is not a free chatbot, it is not a production deployment, it is not compliance assurance, and it is not a substitute for proper discovery work. Thirty days is enough to show you how a bounded assistant behaves on appropriate public or supplied business information that is verified and refined with you. It is not enough to build organisational AI capability, and we are not going to suggest otherwise.

We will also tell you, in writing before anything starts, what we have and have not tested. Our privacy and personal-data handling testing is not yet complete, which is why public demonstrations exclude personal, health and payment information by design. If what you are looking for is a chatbot on your website next week, we are not the right people and we would rather say so now.

At the end of 30 days

The access period ends. Nothing continues automatically, nothing renews, and no invoice appears. We sit down, go through what you found useful and what you did not, and you decide.

Four valid outcomes are to stop, improve the information, continue testing, or design a production solution. All four are acceptable outcomes. The experience does not guarantee that AI will be suitable or that a production system should follow.

See what comes after the trial

Want to see an example first?

Start with one problem

The form takes a couple of minutes. It asks who you are, what kind of organisation you run, and what you would want an assistant to help with. No payment details, and no commitment beyond a conversation.

Start your free 30-day experience

Request a consultation instead