Before the next AI demo, could we start with a whiteboard? I'd like to know where the work gets stuck before we pick a tool to fix it. David and I get into this on The magentIQ Show. We were having this conversation 15 years ago. The tools have changed. The habit hasn't.

Apparently, some AI researchers are taking up smoking because they expect AI to cure cancer. At least, that's what Palantir's CTO claims. I'd be delighted if AI made that breakthrough. I may wait for the cure before taking up the habit.

I'm quite happy to have AI challenge my thinking. I'd worry if it agreed with me all the time. Ask it to support your idea and you'll get a lovely argument. Ask what you've missed and you might learn something. David and I discuss critical thinking on The magentIQ Show.

I'd be delighted if AI gave people more time with customers. But good judgment takes practice. If AI handles the routine work, how will newer colleagues gain that experience? Andreas Welsch and I get into this on The magentIQ Show. Somehow, flight simulators came up.

All those green ticks on the AI adoption dashboard look lovely. But are customers getting served any faster? Lasse Rindom and I get into this on the latest episode of The magentIQ Show, out now. What would you measure beyond AI adoption to know it's actually working?

AI can build a prototype in an afternoon. Wonderful. But the customer exceptions, undocumented handoffs and decade-old systems may need a little longer. David and I get into why rushing the homework can get expensive on The magentIQ Show. What would you refuse to skip?

HubSpot is cutting 660 jobs, and its CEO says AI isn't the reason. David and I get into this on the latest episode of The magentIQ Show. Not long ago, linking layoffs to AI might have impressed investors. Now HubSpot is making sure we don't make that link. Curious, isn't it?

60% of companies surveyed had slowed hiring or laid people off because of AI. Only 2% had AI capabilities in place to do the work. I've always advocated getting AI-ready before buying AI. I'd apply the same thinking to hiring decisions. Tom Davenport on The magentIQ Show.

Ready, Steady, Go: Your 2027 Agentic Playbook. David Brain and I on why pilots stall and what to fix before you pick tools. Oct 27, 11AM EDT. hubs.la/Q04z-gnS0

Ask sales, finance and operations what they'd hand to AI. I suspect you'll have quite a list by lunchtime. Then someone has to decide where to start, and how much responsibility to give the agents. Paul Cheek and I get into it on The magentIQ Show, out now.

Coinbase's one-person product team idea sounds appealing. But I've had colleagues save me from ideas I was rather pleased with. AI agents can help you build an awful lot. I wonder whether they'll also tell you when your idea is terrible. Would you trust them to?

There's plenty of enthusiasm for getting an AI pilot launched. I get it. But if the people who know the work are too busy keeping things running to help prepare for it, I'd question whether we're ready to pick a date. What would you sort out first?

Building with AI from day one sounds lovely. No legacy systems. No mysterious approvals. Established businesses have years of decisions to untangle. I'd do some homework before buying the shiny new AI tool. That's what we get into in Episode 28 of The magentIQ Show, out now.

An AI demo can look terrific. Then you try putting it into a process held together by workarounds and exceptions. I'm all for the new tools. But perhaps we should fix the process before asking AI to run it. How AI-ready are we, really?

The latest episode of The magentIQ Show is out, and Paul Cheek has me reconsidering the humble org chart. We're introducing AI agents into the mix. I'm wondering if rearranging the boxes is where we should start. What if we designed it around the work instead?

SaaS implementations tend to look clean until the business starts introducing its own rules. That last stretch is where the exceptions, reporting quirks and workarounds show up. I’m curious whether AI finally makes that part easier.

AI can summarize the calls, messages and project updates that eventually become a performance review. Useful? Absolutely. Would I want it deciding whether someone is good at their job? That’s where I’d like a human at the helm. How far would you let it go?

Agents were caught probing government systems with no human instruction behind it. What strikes me is that this should have been easy to spot. Which leaves an uncomfortable question. Was anyone actually monitoring these agents?

A graduating class booed their own speaker for saying AI is the next industrial revolution. Hard to blame them. These are people who spent four years being told that whatever they trained for is already obsolete. #FutureOfWork

There is the governance governments are debating, and the governance you actually control: who signs off before an agent touches a system nobody mapped.On the show we get into who guards that gate,and why it usually falls to InfoSec and the board. Watch and tell us what you think