Skip to content

AI & Automation

Adopt AI and automate the busywork, safely. We help you pick the right tools, connect your systems, and put guardrails around your data.
Talk to a consultantAll services

From drafting and support to internal workflows, we help you put AI to work where it actually pays off, then wrap it in the access controls, data handling, and monitoring that keep it safe to use.

The problem is rarely the technology

By the time most businesses ask us about AI, their staff have been using it for months. Not maliciously, and usually to genuinely good effect, but on personal accounts, with no rules about what can be pasted in, and no record of what has been.

That is the actual risk, and it is a governance problem rather than a technical one. Blocking the tools moves the usage somewhere you cannot see. Ignoring it means finding out what went into a chatbot when a client asks.

So we start by establishing what is already happening, without making it a witch hunt. People are usually happy to tell you, provided the question is not framed as an accusation.

What we actually do

Write the policy. Short, specific, and readable: which tools are approved, what categories of information must never go into any of them, and what to do when somebody is not sure. If it needs a legal background to follow, it will not be followed.

Check the terms nobody reads. For each tool you want to use, we establish where data is processed, whether inputs are used for training, how long they are retained, and what the contract actually commits to. The difference between a consumer plan and a business plan is often the whole answer.

Build the automations that pay. Quote preparation, document drafting, support triage, data extraction from PDFs, internal search across your own documents. We look for work that is repetitive, text-heavy, and currently done by somebody expensive, and we measure the before and after.

Set the guardrails. Approved tools available through your existing sign-on, sensible access controls, and logging that lets you answer questions later.

Where we will tell you not to bother

Not every process should be automated, and a lot of AI advice is enthusiasm with an invoice attached.

Automation pays where the volume is real, the input is consistent, and the cost of an occasional error is low or catchable. It does badly where judgement matters, where the process changes constantly, or where an error is expensive and hard to detect. Putting AI into a decision that affects somebody’s money, employment, or safety without a person reviewing the output is a way to industrialise a mistake.

We would rather deliver two automations that save six hours a week and keep working than a broad programme that impresses in a slide deck and quietly gets switched off in March.

Typical timeframe
A policy and pilot in weeks; automation work scoped per workflow
Best for
Teams already using AI informally, and businesses wanting to start deliberately

The engagement

How it runs

Every engagement follows the same shape, so you always know which part you are in and what is coming next.

  1. 1

    Find out what is already happening

    Almost every business we assess is already using AI, just not officially. We start by establishing what staff actually use and what has already been pasted into it, without turning it into a disciplinary exercise.

  2. 2

    Set the rules

    We write a short, readable AI use policy: what is approved, what is banned, and what to do when someone is unsure. Policies people cannot understand get ignored.

  3. 3

    Choose the tools

    We match tools to your actual needs and check the things that matter: where data goes, whether it trains a model, what the retention terms say, and where it is stored.

  4. 4

    Automate the busywork

    We identify the workflows where automation genuinely pays off, build them, and measure the time saved so you can tell whether it worked.

  5. 5

    Put guardrails on

    Access controls, data boundaries, and logging, so the safe path is also the easy one and you can see what is being used.

Common questions

Is it safe to use AI with our business data?

It depends entirely on which tool and which plan. Consumer tiers of many AI services reserve the right to use your inputs for training, and that is the setting most people are on without realising. Business and enterprise tiers generally do not, and contractually commit to it. The gap between those two is where most of the real risk sits, and it is usually fixable with a plan change and a policy rather than a ban.

Our staff are already using ChatGPT. Should we block it?

Blocking rarely works and usually makes things worse, because the usage moves to personal phones and personal accounts where you can see none of it. The better path is to provide an approved tool that is at least as good, be specific about what must never go into any of them, and make the approved route the path of least resistance.

What should never go into an AI tool?

As a starting point: client data you hold under confidentiality, personal information about identifiable people, credentials and keys, unreleased financial information, and anything covered by a contract that restricts disclosure. The policy we write makes this concrete for your business rather than leaving it as a principle.

Will AI actually save us money?

In specific, repetitive, text-heavy work, frequently yes. In general, it is oversold. We would rather find you two workflows where the saving is real and measurable than deliver a strategy document about transformation. If we cannot find a case that pays, we will say so.

Do we need an AI policy if we are small?

If your staff use AI at all, yes, and it does not need to be long. Two pages that people read beats twenty that they do not. The value is mostly in having answered the question before somebody has to guess.

Read more on this

Need help with AI & Automation?

Tell us where you are and what is worrying you. We will tell you honestly what we would do first.

Get in touch