What we build AI & advisory

Advice first, then AI where it earns its place.

Find out where better processes and AI can help, choose a sensible first project and put it to work. We start with how your business runs, the information you have and the result you need.

  • Price agreed up front
  • Scope in writing
  • You own what we build
Interactive example Fictional data
TEAM KNOWLEDGEApproved documents
Find it. Check it. Use it.

Start with a question.

Answer from approved documents

Send your proposed dates to your manager before making travel plans. Your manager checks team cover and confirms the decision.

Team handbookLeave requests · page 8

Check the source before acting. Fixed sample answers.

Choose a question to explore this fixed demonstration.

What you can build

Choose the parts that solve your problem. We agree the features and how they fit together before the build starts.

01

A clear business recommendation

Review the process, customer journey or commercial question. Get advice on what to change and the order to tackle it.

02

A focused first project

Compare possible uses of AI against value, cost, available information and the level of review the work needs.

03

Answers with a source

Help your team find information in approved documents, with references they can check and an honest reply when the answer is not there.

04

Drafts ready for review

Prepare a first pass at a reply, summary or document. Keep the final decision with the person responsible.

05

Help with repeated tasks

Classify requests, extract information and prepare the next action, with agreed permissions and points for human review.

06

Checks before wider use

Test realistic examples, including failures. Agree acceptable results, running costs and a way to pause or roll back the system.

From question to checked answer

An assistant should make answers easy to check. This fixed example shows a response next to its source, with a clear point for a person to review.

Try the interactive examples
  1. 01

    Search only the information the user is allowed to access.

  2. 02

    Show the answer with the relevant source.

  3. 03

    Ask for a person to review uncertain or consequential work.

What's included

Business advisory and AI implementation can be separate engagements. We agree the question, deliverables and price before starting. An AI build also covers data access, running costs, evaluation and human review.

You receive the code we write for you, documentation and a walkthrough. Support and third-party costs are set out before you commit.

See build and support pricing

Common questions

See all questions
Can we start with advice, without committing to a build?

Yes. A review can help you decide what to do next. We agree its scope and price before starting, and the recommendation can include improving a process without new software.

What if AI is not the right answer?

We will explain why and suggest a practical alternative. Sometimes a better form, a straightforward integration or a clearer process does the job.

How is our information handled?

Before a build, we agree which information the system may use, where it can be processed, who can access it and how long it is kept. Sensitive work may need a more restricted setup.

Can the system make decisions on its own?

Only within explicitly agreed boundaries. We define what it can do, what needs approval and when it should stop or ask for help. Important outputs are checked with realistic tests before rollout.

Tell us what you want to make better.

Book a free 30-minute call, or email a few lines about what you need. We'll help you work out where to start.