Subscribe for weekly intelligence on making AI actually work, from the people doing the work.
✓ You're subscribed. See you in your inbox.
I spent six weeks teaching senior product managers and UX designers at Salesforce to work with AI. Here's what actually helped them, and what I'd tell any PM team.
Read the articleNobody rereads a meeting transcript. Claude can turn one into decisions, actions and owners in a couple of minutes. Here's the realistic workflow, the consent basics, and the habit that keeps it accurate.
Read the articleFive days, no coding background required: install, read a file, ship something small, connect a tool, and pick up the judgement habits that make the rest of it work.
Read the articleThere's no sticker price for corporate AI training: it's scoped, not shelved. Here are the factors that move the number and how to compare quotes.
Read the articleThe civil service now gives every person up to 5 days of AI learning a year. Here's how a public sector team turns that into real, hands-on skill — not another click-through module.
Read the articleFrameworks, a DPS, the Procurement Act and two tender sites. Here is how UK public bodies actually buy training in 2026, and where a supplier fits.
Read the articleThe way funded AI and digital training gets commissioned has moved from Whitehall to the region. Here's the plain-English version for West Midlands employers, learners and training providers.
Read the articleYou get up to 5 days of learning a year. Most people never take them. Here's a day-by-day plan for turning them into AI skills that survive contact with your actual job.
Read the articleI sell Claude training, so read this with that in mind. My honest view: most public sector teams shouldn't pick one tool, and the first move is usually to train people properly on the one they already have.
Read the articleYou keep being told to use AI responsibly. Nobody tells you what that looks like on a Tuesday afternoon with a deadline. Here are the three habits that do most of the work, and the checklist to keep by your desk.
Read the articleCopilot lives inside Word, Excel, Outlook and Teams. Claude is a general assistant plus an agent that builds real software. Here's how to pick without a brand debate.
Read the articleClaude can read your spreadsheet, answer questions about it in plain English, and chart what it finds. Here's the realistic workflow, and the one habit that keeps it honest.
Read the articleA realistic AI transformation roadmap runs in quarters: baseline first, then foundations, department activation, scale, and proof of return. Here's the anatomy.
Read the articleAn AI maturity audit scores how your organisation actually uses AI: usage, proficiency, change capacity, self-built workflows, and measured value. Here's how it works.
Read the articleWe're a Claude partner, and we'll say so up front. Both tools are excellent. Here's how to work out which fits your team, and why most companies end up with both.
Read the articleAn AI agent is a model given tools, memory, and a goal, working in a loop with checkpoints. Here's what that means in practice, minus the hype.
Read the articleConsultants diagnose, plan, and build. Trainers change what your people can do. Here's how to tell which one your problem actually calls for, and when you need both.
Read the articleMost ChatGPT training stops at prompting tips. A serious programme covers workspace setup, data controls, role-specific workflows, and the verification habits that prevent expensive mistakes.
Read the articleOn-site brings energy and senior buy-in. Live online wins on cost, reach, and weekly cadence. An honest comparison, plus how to pick for your team.
Read the articleYou need an operator, not a researcher. When the role is worth creating, why internal promotion often beats an external hire, and the interview questions that expose real capability.
Read the articleClaude Code turns Claude into a coding agent that lives in your editor: it reads your project, plans, edits files, and runs commands, with you approving as it goes.
Read the articleOften yes — if it's certified. Most learning and tuition-reimbursement schemes require a credential, not just a course. Here's how to check and what to submit.
Read the articleA certificate doesn't prove you're good at AI — your work does. But the right one removes doubt, unlocks reimbursement, and gives you a reason to actually finish. Here's when it's worth it.
Read the articleNot prompt tricks. The durable skills are judgement, workflow design, and verification — knowing what to hand AI, how to chain it, and when to trust it. Here's the short list.
Read the articleWe built two real products on Claude: Cortex, a brand-knowledge server, and Orbit, a content engine. Here is what actually mattered, and what I would warn you about.
Read the articleMost AI projects that wow in a demo and die in production share one flaw: a clever prompt with no memory and no structure. The ones that last share a shape.
Read the articleOne prompt rarely produces publish-ready work. A multi-agent engine splits the job into specialised steps, the way an editorial team does. Here is how Orbit works.
Read the articleMost AI tools guess your brand because they have no memory of it. A knowledge server like Cortex fixes that by giving every agent the same structured truth over MCP.
Read the articleYes, to a point. Claude lets non-developers build real internal tools, prototypes, and MVPs. The honest limit isn't building the thing; it's knowing when it's safe to rely on it.
Read the articleEvery company has a backlog of small internal tools that were never worth a developer's time. Claude changes that maths, if you also handle the governance.
Read the articleA CRM is mostly structured data and a few workflows, which is exactly what Claude Code is good at. The question isn't whether you can build one, but whether you should.
Read the articleClaude Code makes building a working MVP genuinely fast, even if you're not an engineer. The skill is scoping tight and knowing where a prototype stops being safe to ship.
Read the articleCorporate AI training is hands-on, tailored teaching that gets a company's teams using AI safely and well on their real work. Here's what it covers and why it's different.
Read the articleExecutive AI training isn't a tools workshop. It's the literacy to ask better questions, set the strategy, and use AI visibly enough that the org follows.
Read the articleLawyers have been sanctioned for filing AI-invented case citations. The training that matters teaches confidentiality and verification, not just prompting.
Read the articleMarketing teams adopt AI fastest and feel the brand risk first. Good training is less about generating more and more about keeping it on-brand and true.
Read the articleFinance is high-stakes and regulated, which makes generic AI training useless and good training valuable. Here's what it should cover for a finance team.
Read the articleBuild the capability internally or bring in a provider? It's rarely either/or, here's how to decide, and the hybrid path most teams land on.
Read the articleEvery AI training provider sounds the same on a slide. Here are the questions and red flags that actually separate the ones who change how a team works.
Read the articleAI UX feels new, but the hard questions were answered years ago. Five frameworks (from Google, Microsoft, and decades of HCI research) every team should know.
Read the articleBorrowed from self-driving cars, a levels scale turns "is this agentic?" from a vibe into a question you can answer. Most of what ships today is L3.
Read the articleHalf the arguments teams have about "AI" are really arguments about which layer they mean. Four words fix that, and change how you scope every feature.
Read the articleMIT found 95% of enterprise GenAI pilots show no measurable P&L return. The failure isn't the models — it's how pilots are run.
Read the articleNot "chief prompt officer." The role owns the gap between AI access and AI outcomes — here's the mandate, the org design, and the first 90 days.
Read the articleLicense counts and login charts aren't ROI. A three-layer framework — capability, usage, outcomes — for measuring what AI actually returns.
Read the articleNot the flashiest demo, and not your most painful process. Score candidates on frequency, pain, data readiness, blast radius, and ownership.
Read the articleTo get people experimenting with AI, you have to encourage them to play with it first.
Read the article
Free assessment
Take the free AI Proficiency Predictor: 12 questions, an instant score across Capability, Usage, and Momentum, and the next step that fits your org. 2 minutes, no account.
Start the free assessmentEnter your work email and we'll send you a secure sign-in link.
Request received. We'll email a secure sign-in link to your work email shortly.
