Skip to content
Neurally

AI governance that helps the work.With accountability designed in from the start.

Good governance means knowing, for each decision AI is involved in, who is accountable, what a person still checks, and how the decision can be reviewed and reversed.

Book a free discovery call (opens in a new tab) Where to start, and what it costs

Governance questions usually surface like this.

  • “We have an AI policy. I'm not sure it changes what anyone does.”
  • “If the model gets it wrong, who is accountable? I've had three answers.”
  • “Risk and compliance see things after they're built.”
  • “The board asks how we're governing AI and we describe a committee.”

The cheapest time to govern is before the build

Governance has a reputation for slowing things down, mostly because it tends to arrive late. A team builds something that works, then risk, legal and compliance are asked to approve it, and the design is reopened.

Asked at the start, the same questions are design inputs. Who is accountable when this decision is wrong? What does a person still check? Can the customer have it reviewed? Which data is used, and where does it go? Answering them early shapes a better workflow and removes the late surprise.

Stuart learned this by launching consumer AI inside an FCA-regulated business, before Neurally existed. It is why governance is part of every Neurally engagement.

What governance has to settle.

  1. 01

    Who is accountable for each decision

    Every workflow redesign sets out who is accountable when AI is involved, what stays with a person and what is automated.

  2. 02

    How a decision is reviewed and reversed

    A route for challenge and correction, designed into the workflow before it goes live.

  3. 03

    Which regulation applies

    For financial services and insurance, Consumer Duty outcomes and conduct risk. For everyone, UK GDPR on data use, retention and third-party processing. Where relevant, EU AI Act risk categories.

  4. 04

    What the board sees

    Assumptions, risks and dependencies made as visible as the opportunity, with owners and decision gates in the mobilisation plan.

What tends to go wrong.

  • Governance as a separate workstream

    A framework is written apart from the work it governs. Teams treat it as a hurdle and route around it.

  • Review at the end

    Risk and compliance are asked for a view when the design can no longer change cheaply.

  • Accountability spread across a committee

    When a group is accountable, no person is.

Governance runs through all four stages.

  1. Stage 01

    Decide

    Risk-category exposure and regulatory constraints are flagged while options are still open, and set beside each value case.

  2. Stage 02

    Design

    Accountability, decision rights and review routes are written into each redesigned workflow.

  3. Stage 03

    Deliver

    Pilots have acceptance criteria that cover the controls as well as the performance.

  4. Stage 04

    Embed

    Working norms and training fit your governance, and each result has an owner.

How the four stages work

Where this comes from.

First
ChatGPT app for insurance, and the company's first consumer AI assistant, launched in an FCA and Consumer Duty environmentCompare the Market, before Neurally

These are outcomes from roles Stuart held before founding Neurally. They are not Neurally client results, and the wider commercial figures reflect the work of whole teams, not AI alone.

Governance is not sold separately.

It is a thread through every engagement. The clearest place to see it is the Deep Dive, which includes an AI governance and risk view alongside the operating-model outline and the board plan, in two weeks for £12,500. Boards and non-executives who want continuing independent challenge on governance can use the advisory retainer.

Talk it through first, free (opens in a new tab) Full scope on the services page

Engagement 03

AI Transformation Deep Dive

£12,5002 weeks

  • An operating-model outline: roles, teams, workflows and decision rights
  • The boundaries between people, AI assistance and automation
  • A board-ready 90-day mobilisation plan with owners and decision gates
  • An AI governance and risk view

Also worth a look: CEO, Board & Investor Advisory (quoted after scoping).

Questions we hear.From people in this spot.

Do you provide an AI governance framework?

Not as a standalone product. We build governance into the operating-model design and the mobilisation plan, so that accountability, decision rights and review routes are part of how the work is done.

Does Consumer Duty apply to our use of AI?

If you are an FCA-regulated firm serving retail customers, the outcomes you owe those customers apply whether or not AI is involved. We test designs against Consumer Duty outcomes and conduct risk. Whether and how the rules apply to your firm is a question for your compliance team and legal advisers.

Does the EU AI Act matter to a UK business?

It can, for example where products or services reach the EU. We flag possible risk-category exposure early, while the design can still change. We do not give legal advice, so confirm the position with your counsel.

Are you lawyers or compliance consultants?

No. We design products, workflows and operating models with regulation in view and work alongside your legal, risk and compliance teams. Formal advice stays with them.

How is governance different from the operating model?

The operating model is how work is done and who decides. Governance is the control, accountability and oversight around it. They meet at decision rights, which is why they are designed together.

Related reading

Start with a conversation.Thirty minutes, free.

Tell us what you are trying to decide. If we can help, we will say how and what it would cost. If we can't, we will say that too.

Book a free discovery call (opens in a new tab) Or send a message