Email 1 was about ignoring the circus. Email 3 (coming Friday) is about the AI tools you should try.

But today, we want to talk about the shifts. We debated what 'real progress' looked like this year. Here’s where we landed.

We - Ben, Cien & Henry - spend a lot of time thinking and arguing discussing which AI updates matter and which are just marketing fluff/online hype. Here are the three themes we kept coming back to. And why we think they matter for regular people.

I’ve just written about my reasons to be cheerful at the end of 2025. In my top three was the fact that AI went from being hyped to becoming a real, useful reality - and I’m not just talking about back-and-forth with ChatGPT.

This year I’ve seen a shift with founders, freelancers and small teams setting up what we call 'AI team-mates' to handle content marketing, operations, or finances. These AIs have been basically operating as a team-mate: someone else to delegate tasks to - and off they go and run them!

So that's the big shift this year I've seen: moving from it being something we observed and interacted with a lot to being a reality that we use almost every day, and structured into our work, and helping make a real impact.

(Ben Keene is an Impact Entrepreneur, built an island community in Fiji, runs the world’s best non-fiction book club, raises investment for climate startups, and loves helping people fly with AI)

We kept waiting for a breathless headline announcing AGI, but the real shift happened in the quiet moments. This was the year of "The Crossing." AI tools stopped waiting for us to prompt them and started moving work forward on their own.

Whilst social feeds flooded with "reach content" (or “slop”!) - optimised for speed and repetition - the most valuable work involved slowing down. We learned that AI handles the logistics, but humans must supply the trust, and the final say. The win this year wasn't in replacing the creator, but in building a digital studio that protects the creator’s finite mind from the demand for infinite output.

(Cien Solon is an AI transformation strategist, speaker, and founder of LaunchLemonade, a platform helping non-technical people build AI agents without code, aiming to close the AI literacy gap)

I like to imagine we're in the parable of the frog in boiling water. Except instead of boiling water, it’s.. ever sweeter and more delicious milkshake? Every month in AI means it’s getting better, with more and more delicious choices.

If you want to nerd out and go deep on using the latest AI tools - there’s been something new for you to play with every day. But focusing on the basics - picking a handful of tools, really understanding how they work and what they can do for you, and getting good with them pays off. Gosh it’s hard to avoid the shiny object syndrome though!

I think a great target for 2026 is still what worked this year: just going deep on 2-3 things (I suggest document creation, voice AI for dictation, and simple automations) will save most normal people 5-10 hours per week over the coming year.

(Henry Blanchard is a lecturer in Business & Entrepreneurship, founder of several travel & social enterprises, and advisor to many startups and startup incubators in the UK)

So that’s what we think - AI’s getting good, it can work as a teammate with you, it’s great at content but make sure you have the final say, and don’t get distracted by shiny object syndrome - master the basics to set yourself up for a great 2026.

Tomorrow in the last of this three part series, we’ll share with you the top AI tools and techniques our students on our Get Good with AI programme (and us) have found the most useful, saving them time, effort, headspace - and money!

Until tomorrow,

Ben, Cien, Henry.

P.S Find out more about our next 6 week Get Good with AI courses here:

Cohort 6, starts Tues Jan 13th (online),
Cohort 7, starts Weds Jan 14th (in person, London)
Cohort 8, starts Thurs Jan 23rd (online)