Picture this. Someone in leadership turns to the IT person and says, "Right, what's our AI strategy then?" And the IT person, bless them, reaches for the classic consultant line: "Well, first we need to look at the business's pain points." Noble instinct. Wrong starting point. Here's the uncomfortable truth: the IT person does not know the business's pain points, not really. They know the system outages, the clunky integrations, the ticket backlog. They do not know that the

Picture the scene. A client marches into their solicitor's office clutching twelve pages of "research" their chatbot produced at 2am, complete with three case citations that do not exist and a legal theory that would make a first year paralegal wince. The lawyer now has to read it, fact check it, gently explain why Article 47(b) is not real, and then do the actual legal work anyway. Welcome to the new billable hour: AI Homework Marking. The post that started it On a small b

Every business meeting makes noise. Record it and hand it to AI, and that noise turns into insight, a lighter note taking load, and maybe even a gentle nudge to stop saying "like" quite so much. The real cost of business meetings Knowledge workers sit through 15 to 31 meetings a week, and 71% of senior leaders admit most of them go nowhere. Nationally that works out at roughly 399 billion US dollars a year, about 24 billion lost person hours, and around 25,000 US dollars pe

Construction has never suffered from a shortage of information, if anything it suffers from the opposite. Every project generates a mountain of drawings, RFIs, site reports, emails, meeting minutes, change requests, programmes and compliance records. The challenge isn't creating documentation anymore. It's finding the right piece of information before it becomes tomorrow's problem. That's why most of the exciting headlines about AI miss the point. Construction doesn't nee

Q4 is here and it’s changing conversations (if it’s not, well it should be) Often experiments become business cases and the ideas that follow become budgets. Then the questions shifts from “what could AI Do?” to “What are we actually going to invest in?”. According to Deloitte’s 2026 financial trends research, 23% of finance leaders prioritising AI and advanced technology for FY2027. Meanwhile, 54% are involved in enterprise AI and technology, capital allocation, while

Most bid teams treat the RFP response as a paperwork exercise: assemble the boilerplate, answer every question, hit send. But evaluators don't read that way, and neither does revenue. Procurement research on how evaluators actually work shows a consistent pattern: reviewers sort proposals into a "keep reading" pile or a "reject" pile within the first pages, often the executive summary alone. Busy evaluators read only as much as they need to make that call. If your opening p

Scroll LinkedIn and you'll find plenty of neat four-step diagrams: Idea. Prototype. Integrate. Deploy They look great, it's easy to look at, digest and believe however, just aren't how AI transformation actually happens. What actually stands between an idea and something in production isn't a step. It's four much harder, much less photogenic things. None of them fit neatly into a box, and none of them are optional. The part most AI projects skip Most AI projects start with a

A new era of consultancy has arrived, powered by AI. On our latest engagement, the best design document we produced was never typed, it was spoken. We have just delivered a proof of value for a client operating in a complex, regulated construction environment: an agent that automates the monthly financial review of a major capital programme. Hundreds of live projects, a manual review taking days, and the projects that genuinely needed attention buried deep in spreadsheets. H

AI is now better than checking documents, contracts, proposals and other similar work than you and me. (Good. I never liked doing reviews anyway.) This AI advantage only increases at scale. It can deliver the same level of service 24/7 at a lower cost than you. How many times have you seen stuff leave your organisation and then been horrified to find out what was sent to your customer. Best case is a poor brand impact; worst case is commercial and legal risk. We've built an a

AI only re-uses things that are out there. It's not creative. It can't cope with the demands of a complex ideation process. The history of human creativity is stuffed with so much re-use, it's quite a comical view to adopt. From Bridget Jones's Diary, a copy of Jane Austen, to Roger Sanchez sampling one verse from a 70s song from legendary Toto, the world is full of slightly tweaked copies. There are only so many words, notes and pixel colour values. Creativity is not desirab

From Andrej Karpathy, formerly of leading AI roles in OpenAI and Tesla. I've been saying this will happen for a while now. The charge that has been (reasonably) held against me is that I am selling AI services, so I am biased. I am biased, but that does not mean that a huge change is not a real thing. Software engineering is now a different discipline, with a new set of skills to augment the existing ones. https://x.com/karpathy/status/2015883857489522876

There has been a lot of LinkedIn noise about this AI Skills Hub box of weblinks website delivered to the UK Government by PWC for a cool £4.1m. A good post commenting on this below. I had a look at the new site. It's not a very good website (it's not the worst) but there are so many questions here. How can we do better and spend our public money to get greater value is one obvious one, but the key question, for me, is this: where has our ambition and visionary drive gone? 1.

I just used AI to discover a novel variation in quantum mechanics. This started with a conversation last week with Jeremy Davey. If Einstein had used a 1905 version of the Internet and an LLM could he have created the Theory of Relativity through prompts? The ChatGPT thread below goes on a journey through that topic and on to exploring what a paradigm shift breakthrough actually is. Often, it's a complete reversal of a concrete assumption (the world is flat, the sun orbits us

I have six things to say about this: 1. I’m not sure how a company that sells software to lawyers can triple in valuation to $6B in four months. 2025 revenue was just $23M. This is not unique to Legora, there are other AI SaaS tools that do “one thing” with these odd valuations. It only makes sense to the previous round of investors. 2. AI SaaS is not the future. SaaS is not dead but it is starting to feel a little tired in the morning. Custom AI solutions on a private and go

I was at the Microsoft AI Tour last week. Microsoft exists to sell products. I worked there many years ago in their industry team and the license always won over the solution. It’s in the DNA. They do make some great products. We are a top-tier Microsoft AI partner and we deploy them in our engagements. But AI is not about software products. It's so far away from what most businesses need that its like trying to sell them a chocolate eclair to hammer in a nail. AI transformat

Great article in The Times this weekend. Our country needs to invest more in supporting AI startups, amongst other future tech. If you want to support a U.K. business with your AI plans then drop me a line. Ian Hogarth We need to back ‘patriotic’ winners to build the UK’s first £1tn tech giant https://www.thetimes.com/article/a67e6cf2-eb60-42bb-a30e-8691d2a7c39d?shareToken=449eb80e7717538f6df4718b65416603

There is lots of talk about AI taking jobs. This might happen. It might happen quickly or slowly. New categories of jobs will appear, roles will change task mix and some tasks will completely disappear. The net position will likely not alter too much. This is what history tells us. When Google first bounced on to the scene, there were no SEO agencies. And now there are. With AI they become GEO agencies or they perish. And so the cycle continues. The best defence? Network wide

Clarity and repetition of key themes are fundamental tenets of messaging. If you asked your team these questions, do you think you would get consistent answers from them? 1. What does our business do? 2. What is the top priority for our business? 3. What is your role in that and what are your goals for this week/quarter? In my experience people find it hard to move their thinking to 1 and 2 when they spend so much time in one detailed element of 3. The context switch is jarri

Let's get experimental. It's very easy to get stuck in the mindset of we've always done it this way and this is particularly true when your previous efforts have been successful. You may notice that when you're not being successful there is a tendency to try lots of new things, perhaps somewhat in desperation. Neither of these is optimal, of course. In either state it is important to try new things in a controlled way that allows you to test a hypothesis and evaluate what var

Are you frustrated that your teams don’t want to change or try new ideas? I don’t think it’s possible to enforce change in another human. Change is an internal process, not an external one. So what can you do to help? You can give people a new input: a piece of data that could change their perspective. As an example, let's say you know you are overweight and you want to do something about it, but you don't persist with diets and lifestyle choices. I've been through this mysel