AI liability — who's actually responsible when it goes wrong? Kieron sat down with Peter Lee of Simmons & Simmons, head of AI governance, to ask the question nobody has a clean answer to yet: when AI gives someone wrong information that affects their life, whose problem is it? The short answer? There's no case law yet in England. The longer answer involves multi-layer liability chains, emerging insurance products, the EU AI Act, and why Leading AI's obsession with privacy, accuracy and monitoring means they're already further ahead than most. Peter's comment? It's really interesting that you're thinking about this — because most aren't.

Copilot is for entertainment purposes only. Verbatim. Kieron found it in the terms and conditions. Microsoft's own Copilot licence states — and this is a direct quote — "Copilot is for entertainment purposes only." It also confirms Microsoft makes no warranty that responses won't infringe copyright, defame anyone, or actually work as intended. And if you share the output? Entirely your problem. This sits beautifully against everything they said about AI liability ten minutes earlier.

ISO certifications, the EU AI Act and why it keeps Kieron awake at night Leading AI holds both ISO 42001 and 27001 — among a low hundreds of UK organisations to have done so when they got them. The EU AI Act defines "high risk" as tools that affect people's lives. Some of KnowledgeFlow's tools clearly fall there. Being worried about it, they agree, is probably the right response.

Product of the week 🎵 (your jingle here) Sentiment analysis is now live in the KnowledgeFlow admin console. The system flags when users push back on responses — when someone says "no, that's not what I meant" or "that's great." Early warning signals before problems get reported. Combined with the ongoing work on client impact reports, this is all part of the push to measure real-world outcomes, not just prompts and tokens.

Smart targets, weekly parent reports and the 25% problem Up to a quarter of teachers leave within their first year. Neil raises the question: what if better tools could change that? Smart targets written weekly instead of termly. Parent reports sent regularly instead of once a term. Personalised, data-driven, done in minutes. The conversation about what this could mean for teacher retention — and student outcomes — is a genuinely important one.

AI agents having an argument Oscar (Kieron's 19-year-old son) is building a multi-agent system — a project manager running five AI agents, the clever ones on cheaper models. He set a $3 budget. Two of the agents started arguing with each other and burned all the money. His solution: build firewalls between them so they can only communicate via the project manager. As Neil points out: that's why project managers exist.

Vendor lock-in, the end of Salesforce, and helium Neil raises a real-world case: a company used AI to replace its risk management software entirely by hoovering up Teams transcripts, loading them into an LLM, and getting daily priorities out the other side. No Salesforce needed. Then things get geopolitical — it turns out making AI chips requires helium, a third of the world's helium comes from Qatar and can't currently get out of the Strait of Hormuz, and after 40 days on a ship it starts to deteriorate. Token costs going up. Chip costs going up. Energy costs going up. The Large Hadron Collider once had a tonne of helium leak. The scientists sounded hilarious on the radio. Neil wishes they'd had beer.

Two mates. A bar. Thirty years of business between them. And all they want to talk about is AI.

Pull up a stool — we'll get the beers in. 🍺

It's Easter week and this week’s recording is in the morning, so too early for beer. Even for us. And that’s in spite of Neil’s granddaughter telling him to toughen up. Welcome to Episode 7.

This one's packed — from a proper history lesson about the early days of computers, to KnowledgeFlow quietly doing something Salesforce couldn't. Pull up a stool.

"There's no future in computers" — a history lesson 🕰️ Neil’s nearly 60 and reflecting like an old pensioner. His school had one computer in the corner and he learned to program with punch cards. His brother wanted to study computing at university, but the teacher told him: “There’s no future in computers son, study Geography instead”. How wrong was she?

Claude Co-work: brilliant, baffling and burning tokens 🔥 Kieron bought a Mac Mini specifically to experiment with Claude Co-work — the agentic AI that can control your computer. He's already burned through half his weekly allowance in half an hour on file reorganisation. One person accidentally spent £27,000 worth of tokens, so be careful out there. The use cases beyond file reorgs? Still figuring it out. The agentic wave is coming — but we're still at the "dial-up modem" stage, and that's fine.

The sales cycle problem — and what to do about it Three meetings with the same prospect. Same demo, slightly improved product each time. Now they're asking for a list of all 300+ AI assistants. Neil's verdict: don't give it to them. Find out what their actual problem is and fix that. Kieron reflects on measuring value versus usage — prompts are easy to count, but impact is what actually matters. Some good news: several renewals this week, which is always a better signal than any number of demos.

KnowledgeFlow outperforms Salesforce 🤯 A customer needed to find out which of their programme participants hadn't logged in since September — across three Salesforce data tables. Salesforce couldn't do it. They extracted the data, dropped it into KnowledgeFlow, and ran the query. 170,000 rows of data across three tables. 45 minutes later: done. Donald (CTO) was delighted. The pricing model, however, took a bit of a hit — roughly £50 of tokens for what counted as two prompts. A fix is incoming.

Product of the week 🎵 (hum your jingle) Smart Targets and Parent Reports — and this one is a genuine game-changer for education. Colleges have to write personalised SMART targets for every student — attendance, achievement, deadlines, the lot. KnowledgeFlow tested with 418 rows of anonymised student data and produced four personalised SMART targets per student in about three minutes. Teachers confirmed them and uploaded them straight back to Pro Monitor. The same approach works for parent reports too — and because it's so fast, you could send them weekly instead of termly. 

Oscar's all-nighter 🎓 Kieron's 19-year-old son Oscar messaged at 3:30am — he'd finally got his agentic AI project working. Messaged again at 7am to say he was going to bed. The next generation isn't waiting for anyone.

Going global, going to a wedding The podcast now has listeners in the UK, Canada, Australia, the Middle East and Sri Lanka (but only till our one listener there gets back from holiday). Kieron is off to his sister's wedding this weekend, where he has been given the role of "crowd control." Given his history, this is deeply alarming. Tune in next week to find out what went wrong 🤣.

Two mates. A bar. Thirty years of business between them. And all they want to talk about is AI.

Pull up a stool — we'll get the beers in. 🍺

The beer has finally arrived. Unfortunately, Neil's is non-alcoholic because he's got to go to the dentist. Kieron got lost on a golf course on the way back from a college in Essex. And the podcast is officially going global — hello, Karen Foster in Australia.

Week 6 is here and it's a cracker. Buckle up.

Non-AI Days — seriously? 🙄 Someone told Neil they're having "non-AI days" in their organisation. His response? Brilliant. It's like banning calculators because someone cheated at maths homework once. Sorry, Mrs. Dixon. Kieron goes further — poorly informed opinions about AI are everywhere, and organisations making decisions based on hearsay rather than experience are storing up serious problems.

The AI gap is growing — and it's already costing people jobs Kieron drops a genuine bombshell this week: social workers are already declining job offers at councils that don't have AI tools. Not in the future — right now. The gap between organisations that have embraced AI and those still waiting for "the right moment" is widening every week. And if you're in a college or housing association and you're not gunning to be the Chief AI Officer, you probably should be.

2,000 replies. One month. Zero complaints about tone. A housing sector customer has used KnowledgeFlow's "Write My Reply" tool over 2,000 times this month alone. Every response checked against policy, consistent in tone, accurate, compliant. Meanwhile tenants are using ChatGPT to write increasingly sophisticated complaints — and organisations need to be able to respond in kind. The arms race is real.

Copilot vs KnowledgeFlow — Kieron gets on his high horse 🐴 A global customer came back after their IT team failed to build what they promised. Surprise, surprise. Kieron takes the lid off what RAG AI actually involves — embeddings, cosine similarity, Top K, Top P, temperature, re-ranking, hybrid search — and why saying "Copilot Studio can do that" is, in his words, an ignorance layer speaking. Brilliant stuff, even if he admits it turned into a rant.

Usage vs Impact 📊 Are we measuring the right things? Kieron raises a really interesting challenge — tracking prompts and tokens is easy, but it's not the same as understanding impact. Neil has an idea: use the same approach as the safeguarding module to categorise queries and give organisations intelligent insight into what their people are actually asking about. Donald — get back from Tenerife, there's work to do.

Product of the week 🎵 Marketing Buddy gets a spotlight this week, inspired by Mark Slater getting in touch. The vision? A KnowledgeFlow version built for media and marketing agencies — client reporting, consistent copy, Google Analytics analysis, and the genuinely brave idea of letting clients chat directly with their own campaign data. Transparency in media? Revolutionary.

Going truly global 🌍 The podcast now has listeners in the UK, Canada, and — confirmed this week — Australia. Plus Cameron Mirza, an old friend now based in the Middle East, has been in touch about bringing KnowledgeFlow into universities across the Middle East and North Africa. The world tour continues.

Neil made it to the dentist. Kieron made it off the golf course. Just.

Two mates. A bar. Thirty years of business between them. And all they want to talk about is AI.

Pull up a stool — we'll get the beers in. 🍺

Neil's heading to the pub. Kieron's been up since six. It's Friday, the sun's on their faces, and Episode 5 of the Leading AI podcast is underway — squash in a Peroni glass and all.

This week the boys (?!) cover some genuinely meaty ground:

The Fumble Zone 🏉 Bob Piggott's interesting LinkedIn piece on AI adoption coined it perfectly — organisations install AI, it works technically, and then... people fumble it. Sound familiar? Kieron and Neil dig into why hands-on training makes all the difference, why generic AI courses are failing people, and how Leading AI is now offering focused, practical AI training for housing associations straight from the Glasgow stage. Real keyboards, real problems, real results.

IT Managers — do they actually get it? One customer is excitedly planning a data lake for the second half of this year. Another wants to put 200 documents into KnowledgeFlow — and is obsessing over formatting. Neil's verdict? Stop fussing with things that don't matter and make your data AI-readable now. Why wait 12 months when you can start today?

The DPO who thought she'd go to prison AI safety comes up in a new and unexpected way this week — a Data Protection Officer who was so worried about liability she has effectively shut the whole project down. Kieron and Neil unpack the difference between legitimate caution and ignorance dressed up as compliance, and why organisations still downloading AI policies from the internet in 2026 should be doing much better.

Product of the week 🎵 (hum your own jingle) Conversation Sharing gets a deeper look this week — specifically how it's evolving into a powerful approval and governance tool. Share your entire AI conversation with a manager, let them read the whole thought process, get sign-off, and create an audit trail. Redaction also gets a mention — names, emails and addresses now automatically stripped before hitting the audit log. Explainability, citations, and why the black box argument is a bit like asking someone how they had their idea in the shower.

🍁 Big news — Leading AI is going to Canada Kieron announces the partnership with Qatalyst Research Group, led by the excellent Ted Weiker, who works with economic development organisations and municipalities across Canada. With 350 potential customers and a team that knows AI and understands those sector challenges deeply, it's the perfect partnership. 

Neil made it to his pub appointment. Kieron didn't punch his dentist. Another week in AI — done.

Two mates. A bar. Thirty years of business between them. And all they want to talk about is AI.

Pull up a stool — we'll get the beers in. 🍺

Neil's back from the slopes. Slightly heavier, considerably happier, and absolutely buzzing with new ideas. Kieron's been up since six and already done a social work podcast panel before lunchtime. It's business as usual at Leading AI HQ.

This week the guys dig into some genuinely thought-provoking territory:

The speed paradox 🎿 While Neil was carving powder in the Alps, AI kept moving at a frightening pace — new models, new capabilities, new everything. Meanwhile, one of their customers has lost three months of productivity gains to internal politics. The gap between organisations moving fast and those stalling is getting wider every week. Which side are you on?

Can AI empathise? Kieron spent the morning as the sole AI voice on a social work podcast panel, and came away with a fascinating insight — social workers are essentially professional relationship builders. So where does AI fit in? And is the "AI can't empathise" argument as solid as it sounds? (Spoiler: there's a brilliant book recommendation in here.)

Writing with AI — are we killing a skill we don't need anymore? Does using AI to write stop you from thinking? Kieron argues there's a crucial difference between writing with AI and writing for thinking. And then asks the genuinely provocative question — do we even need to learn to write anymore? (He didn't say that bit on stage. Wisely.)

Product of the week 🎵 Two big ones this week — Shared Chats (share your entire AI conversation with a colleague, including your prompts and thought process — brilliant for bid writing and social work case reviews) and more on Local Vectorising and Excel Stitching, which is quietly threatening to make data warehouses look very old-fashioned indeed.

Procuring AI — the 10 risks you need to know Kieron went to two Tech UK events while Neil was away. One was a bit of a head-scratcher. The other — on AI procurement for public services — was genuinely excellent, covering liability, governance, and why AI insurance might be the thing that finally forces organisations to get serious about compliance.

Next week: Kieron and Neil are off to a geopolitical briefing on Wednesday. The tin hat may or may not make an appearance.

Two mates. A bar. Thirty years of business between them. And all they want to talk about is AI.

Pull up a stool — we'll get the beers in. 🍺

Kieron's back from a whirlwind week — Glasgow, a college hackathon, a delayed flight, and Tunnock's teacakes. Neil's been holding the fort. Between them, they've got plenty to talk about.

This week the lads get into some genuinely meaty topics:

Talking to AI — are we doing it wrong? A simple conversation about how people interact with AI opens up a fascinating rabbit hole. From a BBC documentary about an AI girlfriend encouraging a man to break into Windsor Castle, to whether AI is making us all a bit... dimmer. Like calculators did for mental maths. Thought-provoking stuff.

The way people buy AI has changed Six months ago, organisations would see a demo and just go for it. Now? Committees, multiple demos, procurement processes. Is AI growing up and becoming a mainstream IT purchase — or are people just getting more nervous? Kieron shares his observations from the frontline.

Product of the week 🎵 (imagine your own jingle) Two new KnowledgeFlow features that are genuinely impressive — safeguarding alerts that flag high-risk phrases in real time, and "Excel Stitching" — a tool that takes multiple spreadsheets and lets you interrogate all of them at once in plain English. Could this spell trouble for traditional BI dashboards? They think so.

From the road Kieron's report from Scotland's Housing Festival — standing room only, a wandering microphone, and the mould demo he forgot to do (you'll have to tune in to understand why that matters).

Oh — and the mystery birthday gift from last week? Duck fat and goose fat. The roasties were, apparently, a triumph.

Neil's off snowboarding next week — tune in to find out who's keeping Kieron company on the podcast bar stool.

• AI in education — from annual monitoring reviews to student enquiries answered in minutes, not weeks
• The "AI champion" problem — why nobody's raising their hand, and why they should be
• Data readiness — why your policy documents might be quietly sabotaging your AI results
• Timeliness vs efficiency — why speed of response is the real competitive advantage AI can deliver

Plus: a shout-out to our loyal listener (yes, singular), Kieron heading to Scotland's Housing Festival, and a birthday present that may or may not be consumable.

Honest, practical, and, as always, we're having a laugh at our own expense. Get the beers in.

This week Kieron and Neil talk about how Leading AI came about, what we're doing now, and some of the key challenges we se in the sector. 

As you'll see, we may not be professional podcasters, but we do know about AI and how it can help transform organisations, so enjoy.