The real AI edge? Workflow, not wizardry + Your free prompt

Friends,

Your weekly AI briefing is here — designed to help you respond to AI, not react to the noise. No curveballs. No chaos. Just clarity.

🚀 Congratulations! You’re reading one of the Best AI newsletters in the world!
Last week, The AI Optimist was voted one of the top AI newsletters globally by Zapier — Next week we’re launching our Youtube series bringing AI Optimism to a screen near you….

👉 Share AI Optimism with your tribe.
This is the perfect time to pass on pragmatic optimism to a leader who could use a weekly nudge of signal over noise. 🤓

📰 This was the week that was...

This week, OpenAI and Anthropic shared in-depth data on how people actually use ChatGPT and Claude. But don’t let the surface stats distract you - the real story is in what people do with the tools.

🔑 Key insight: AI success isn't about having the flashiest model - it's about how effectively your team uses it, and whether it's truly wired into the way you work.

OpenAI’s report shows ChatGPT is becoming the new Google - 73% of prompts are non-work related, with only 4.2% involving code. It’s a powerful surface for curiosity, advice, and writing - but not yet a workhorse.

Claude is the opposite. Anthropic’s index reveals it’s being used for work - over 36% of prompts are coding-related, with admin, finance and writing making up another 40%. Their API data shows 77% of enterprise use is automation-focused.

📊 The Humaine Leaderboard makes the same point differently. Even when models perform similarly on benchmarks, human users rate them very differently. Why? Because 80% of quality still depends on how they're used - not how they're built.

💡The real takeaway: We’re entering a phase where the winners won’t be those with the most advanced tools - but those who’ve mastered the art of applying them. Success depends on smart processes, good design, and real integration.

Let’s get into it.

🔥 There is no Urgent Priority

Nothing this week requires action within 90 days. No major security events, legal rulings, or sudden vendor shifts. Let’s enjoy that.

🎯 Strategic Insight: Outputs over benchmarks

What’s the Tension?
Should you optimise for better tools, or better workflows?

Optimistic insight:
Claude and ChatGPT tell two very different usage stories. But both confirm that how a tool is used matters more than what it technically can do.

What’s shifting:
Forward-thinking teams are building closed-loop systems. Inputs come from your docs, tools like NotebookLM extract and organise what matters, and tools like Gamma format and publish in real time - on-brand and on message.

Why it matters now:
Building your context today gives you an advantage tomorrow.

✅ Takeaway:
Don't chase the model upgrades. Build AI into the steps where your team actually makes decisions - and treat that process like the product it is.

🧐 A Free Prompt from AI Night School

A prompt for building better workflows with AI | AI Night School

This is the prompt I use when I want to break a task into multiple steps, especially if I’m working with large volumes of information, or if I want the tool to behave differently at each stage (brainstorming → decision-making → writing).

Why it matters: When you ask an AI to do too much at once, it tends to shortcut the process. Multi-step prompts force it to think like a workflow - and that produces more consistent, higher-quality results.

To make it easier, I created the CHAIN framework:

C — Clarify the Goal
H — Highlight the Inputs
A — Assemble the Steps
I — Indicate Tools & Roles
N — Nest the Output

Think of it as: Chain the steps - don’t lump them.

👉 Get the free prompt

🤓 Geek-Out Stories

Humaine Leaderboard: Human-rated models, not just benchmarks

The Humaine Leaderboard ranks AI models based on how people experience them across diverse user groups.

Why it matters: It confirms what we keep saying - operator skill trumps model architecture. A mediocre model with a great prompt beats a top model used badly.

NotebookLM: quietly becoming a dark horse
While bigger names jostle for attention, NotebookLM keeps shipping genuinely useful features. This month: Video Overviews, quizzes, and flash cards based on your sources.

Why it matters: This tool might quietly win the AI interface race - especially for teams who value clarity, not chaos.

Gamma 3.0 brings brand control to AI-generated decks
If you churned out of Gamma (The alternative presentation tool), you weren’t alone. But Gamma 3.0 may just bring you back - with a public API, agentic workflows, and team-brand controls finally in place.

Why it matters: On-brand AI output has been surprisingly slow. This is the start of AI for the presentation layer.

🎨 Weekend Playground

Weekend watch: Dan Kelsall’s “Bro-hemian Rapsody”
A funny, slightly brutal roast of scammers, fake gurus, and the mess some people call AI strategy. Watch it here

Why it matters: As well as a great example of human creativity meeting AI production, it’s funny and that’s good for your health.

🧠 AI at Work

We’ve started tagging our internal prompt logs this week using just three categories: Asking, Doing, and Expressing. It’s already changed how we structure tasks and handovers - less “chatty”, more outcomes.

Want to try it? Set up a Slack channel or Notion table and review team prompts every Friday. It’s one of the quickest ways to spot where your AI value is (and isn’t).

📢 Now really is the time to share the Optimism!
If The AI Optimist helps you think more clearly, forward it to someone else navigating the shift.

You’ve got the perfect reason to shout about it - you’re one of the very first subscribers to one of the world’s top AI newsletters 🎉

And if it’s not quite landing? Hit reply. I read every message.

Stay strategic, stay generous.

Hugo & Ben