AI Accountability for UK Businesses

Welcome to this issue of The AI Brief!

The FCA just confirmed it won’t write new AI rules, because the old ones already make it clear that you, not the software, are accountable for what AI gets wrong. This issue covers what that means for your business, breaks down the £200m just announced for AI adoption and skills and gives you an honest look at whether Bitrix24‘s all-in-one platform earns its place in your stack.

Prompt Engineer

If something goes wrong with an AI-assisted decision in your business, you need to know who signed off on it. This prompt builds a simple checklist showing which decisions use AI, who’s responsible and what to do if it goes wrong, so you have something to point to if anyone asks.

“Act as a specialist AI governance advisor for UK small businesses. I want to create an internal oversight checklist showing who is responsible for AI-assisted decisions in my business, so I have a clear record I can point to if something goes wrong.

Context:

  • My business: [Insert brief description, e.g. 12-person marketing agency]
  • Areas where AI is currently used: [Insert list, e.g. content drafting, customer service chatbot, CV screening, financial forecasting]
  • Team structure: [Insert role titles only, not personal names, e.g. Marketing Manager, Ops Lead, Founder]

For each area where AI is used, please set out:

  • Decision type: What is being decided or produced with AI’s help?
  • AI’s role: Does it fully automate this, assist a human, or just draft for review?
  • Responsible role: Which job title is accountable for the final decision?
  • Human check required: What does a human need to verify before it goes ahead?
  • Review frequency: How often should this be checked once it’s running (e.g. weekly, monthly)?
  • If it goes wrong: What is the first step to take?

Format: Present this as a table with one row per AI use case, followed by a short summary flagging which areas carry the highest risk and should be reviewed first.”

AI Updates Roundup

This week UK regulators and government laid out where the money is going and where the responsibility sits, as AI adoption moves from ambition to obligation.

Overview

  • The FCA ruled out new AI rules for financial firms.
  • The government unveiled over £200m to boost AI adoption and skills.
  • The ICO confirmed new agentic AI guidance is coming this year.

The FCA has confirmed it won’t bring in new AI rules for financial firms, leaning on existing rules that hold senior managers personally accountable instead.

  • So what? If you’re FCA-regulated, this confirms senior staff stay personally liable for AI decisions, so it’s worth reviewing oversight now.

The government has announced more than £200m in funding to help British businesses adopt AI and train staff in AI skills.

  • So what? You might want to keep an eye on the free AI Skills Hub courses, since they’re expanding and cost nothing to try.

The ICO will publish a new code of practice and guidance on AI agents, tools that act on your behalf, later this year.

  • So what? This is the kind of guidance that usually answers the practical questions everyone’s been guessing at.

“Britain’s future will be shaped by the choices we make now.”

Who’s Liable When AI Gets It Wrong? ⚖️

If an AI tool gets something wrong in your business, who’s actually on the hook? It’s worth having a clear answer before you need one.

Overview

  • Responsibility sits with you, not the software.
  • If AI shapes a decision in your business, you need to be able to explain why.
  • Pick one AI-assisted process this week, name who owns it and write down the reasoning behind it.

That’s the principle behind a decision the FCA just made. It’s confirmed it won’t bring in new AI rules for financial firms, because the old rules already cover it: whoever signs off on a decision is responsible for it, AI-assisted or not.

You don’t need to work in financial services for that to matter. If an AI tool drafts a contract, screens a job application or replies to a customer and gets it wrong, the responsibility doesn’t disappear into the software. It lands on whoever let it run unchecked.

What to do this week 💡

Pick one process where AI currently shapes an outcome, whether that’s pricing, hiring, customer responses or anything financial. Then work through the following:

  • Name the owner. Who is the person responsible if this goes wrong? No, it can’t be ‘the team.’
  • Check what the AI is actually doing. Is it drafting something a human reviews or is it making the final call unsupervised?
  • Write down the reasoning, not just the output. If a client, a bank or a court asked why a decision was made, could you explain it in plain English? Once you can, use that reasoning to spot where a check is missing. If you can’t easily explain a step, that’s usually where to add a human review or a rule the AI has to follow before it acts. This matters because it catches mistakes before they reach a customer, not just after.
  • Set a review point. Decide now how often this process gets checked, so you’re not leaving it until after something goes wrong.
  • Document it. A short note in a shared file is enough. The point is that it exists.

If you get to the end of that list and can’t answer one of them, that’s the gap to close first, not a reason to stop using AI altogether.

Quick Review: Bitrix24🗂️

Bitrix24 is an all-in-one platform that bundles CRM, project management and team communication into a single subscription. For UK SMEs juggling separate tools for sales, tasks and chat, it offers a way to bring everything under one roof.

Key Takeaways

What’s Good ✅:

  • Free plan includes unlimited users, which is rare for a CRM at this price point
  • Built-in automation handles repetitive sales and task work without extra setup
  • Over 300 integrations including QuickBooks and PayPal

What’s Less Good ❌:

  • Interface is cluttered and takes time to get comfortable with
  • Live support is often slow to respond when something goes wrong
  • Best run by someone with the time or technical confidence to manage it properly

Impact on Your Business 📈: If you’re currently paying for separate CRM, project management and chat tools, Bitrix24 lets you fold them into one system. It suits a growing business with someone in-house willing to act as the platform admin and iron out the setup.

Financial ROI 💰: The unlimited free plan removes the cost of testing a CRM before you commit. Bundling automation and project tools into one subscription can also trim what you’d otherwise spend on several standalone apps, though the learning curve may eat into those savings in the first few weeks.

Price Point 💳: A free plan with unlimited users is available. Paid plans start at around £69 per month for the Basic tier covering five users, with lower monthly rates if you commit to annual billing.

Upcoming Events & Conferences 📅

A full day on AI in operations, risk modelling and customer experience, free for financial-institution professionals with tickets from £599 (+VAT) for everyone else.

A half-day session for operators, risk leads and suppliers across telecoms, energy, transport, water and data centres on AI assurance in critical systems.

A vetted enterprise AI summit for CIOs, CDOs and CTOs weighing up AI investment, governance and returns.

For C-suite, transformation and operations leaders moving AI pilots into governed everyday business workflows.

A conference for senior public sector leaders and technology suppliers exploring how AI is being deployed across government services, useful if you work with public sector clients.

A session for business analysts and operations teams redesigning processes around AI agents, available in person or online.

Training & Skills Development 🎓

A hands-on day for founders and SME leaders ready to move from AI experimentation to actually putting it to work.