AI in coaching, visual insights with Nano Banana Pro, behavioral data science, I Love Lucy, and more

📖In this issue

• Framework: AI in Coaching (with Nano Banana Pro)

• Humans + AI update: AI is not good for reducing staff, occasional not daily use predominates, accelerating science and scientists.

• From Humans + AI Explorers Community: I Love Lucy in the chocolate factory, collective intelligence with network mapping, avoiding the promotion trap, and more

• Humans + AI Podcast: Ganna Pogrebna on behavioural data science, machine bias, digital twins vs digital shadows, and stakeholder simulations.

💡Framework: AI in Coaching

I attended a fabulous panel discussion organized by Josie Gibson of Catalyst Network on AI in Coaching, with Jax NiCarthaigh, Simon Buckminster-Shum, and Paul Garcia. It was a deep and rich discussion, which was summarized in the Humans + AI Community.

In other news this week, Google has further improved the state-of-the-art in LLMs with its release of Gemini Pro 3, which I am finding extremely good, as well as Nano Banana Pro, an improved version of its already best-in-class Nano Banana model. The model is exceptionally good at infographics, diagrams, and schematics.

Below is what it did (with a little pre-framing in Gemini Pro 3) in summarizing the key themes of the AI in coaching conversation. Pretty good!

🧠🤖Humans + AI

Leader expectations of AI reducing staff are being disappointed

A new ISG report on the State of Enterprise AI Adoption shows that relative to expectations AI is not enabling staff reductions, but it is improving products, services, and experience.

State of Enterprise AI Adoption 2025
Key insights and discussion

Most users of AI are occasional, not daily

Benedict Evans famous annual presentation, this year on “AI is eating the world”, points to the big gap between daily and irregular usage of LLMs, questioning why OpenAI is releasing weekly usage stats rather than daily, as most tech companies do (along with great analysis of the state of the broader ecosystem within AI is set).

AI eats the world
Summary and key insights

AI science accelerates discovery and augments scientists

In a significant advance in AI-accelerated science, Kosmos automates data-driven scientific discovery, running iterative cycles of literature search, hypothesis generation, data analysis, and validation. Despite the “autonomous” framing it does appear to come from a Humans + AI framing.

Kosmos: An AI Scientist for Autonomous Discovery
Summary and key insights

🌐From Humans + AI Explorers Community

Lots going on in the Humans + AI Explorers Community.

One of the gems shared in the community from the AI in coaching panel discussion mentioned above was Jax NiCartaigh’s mention of the segment from I Love Lucy in the chocolate factory as a metaphor for the state of AI, delightful!

Lee Christopher Grant shared a detailed description of Stuart French’s methodogy for bringing collective intelligence to bear using network mapping, in this case in framing the future of knowledge management.

Lavinia Iosub provided sharp insights on how AI is finally freeing specialists from the promotion trap and the tyranny of becoming mediocre managers.

There was deep discussion and engagement on the topic of Prompts to guide LLMs towards augmented cognition.

And far more, with a host of upcoming events!

🎙️This week’s podcast episode

Why you should listen

Behavioral data science provides one of the most powerful ways to guide AI and gain value, as well as to shape ethical perspectives. Ganna Pogrebna is a world leader in the field and excellent communicator of why this field matters and how to apply it.