Greetings, fellow humans (and possibly advanced AI entities lurking in our subscriber list). This week, we're diving into the deep end of the GenAI pool, where chatbots are passing Turing tests better than we are, and dolphins are apparently having more complex conversations than our last team meeting. Will our robot overlords be fluent in dolphin by next week? Stay tuned, and try not to let GPT-4.5 make you question your own existence.
GenAI News Podcast
Don't have time to read this newsletter?
Business News & Market Insights
1. Therapy / Companionship – up from 2nd place 2.Organizing My Life – new entry at 2nd place 3. Finding Purpose – new entry at 3rd place 4. Improving Learning – up from 8th place 5. Generating Code – up from 47th place 6. Generating Ideas – down from 1st place 7. Entertainment and Fun – same position as last year (6th) 8. Enhancing Code (for professionals) – up from 19th place 9. Creativity – up from 27th place 10. Healthier Living – up from 75th place
Note that only few of them are complete process use cases: 31% of top use cases fall under personal and professional support, (+14% vs. 2024), 15% were technical assistance and troubleshooting (-8% vs. 2024). While new tendencies such as Vibe coding might become a bottleneck, GenAI products help reduce mental load, guide action in daily lives. Productivity gains will certainly emerge from a change in employee attitudes, as they will be in a better position to carry out existing tasks.
> Major deals, Partnerships & Innovations
New Models & Innovations
> Image, Video and Other Multimodal Models
Thoughts of the week by Hanan Ouazan
Managing Partner & Global Lead AI Acceleration
When Conversations Become Channels
Brands have long envied the troves of first‑party data sitting inside retailer systems. The average shopper enrolls in roughly 13 loyalty programmes but stays active in barely half of them, so the “relationship” often stops at the cash register. A new inflection point is emerging. Traffic routed from generative‑AI chatbots to US retail sites exploded by 1,950% last Cyber Monday, yet those visits still represent well under one percent of global search volume today. Google’s own AI Overviews now surface in fewer than 15% of logged‑in queries. That puts conversational discovery in the “one‑to‑low‑single‑digits” range – small enough to be ignored in quarterly dashboards, large enough to bend the arc of search behavior. Two open standards suggest the bend will accelerate. Anthropic’s Model Context Protocol (MCP) turns a brand’s product catalogue or order system into a plug‑and‑play socket for any LLM, eliminating the need for brittle screen‑scraping or custom APIs. Google’s new Agent‑to‑Agent (A2A) protocol adds the missing lingua franca that lets those assistants negotiate tasks – “check stock, apply my perks, ship express” – across vendors and ecosystems. When both layers converge, a shopper could research, decide and purchase inside the same chat thread, with the retailer touching only fulfillment. That possibility doesn’t pit brands against retailers so much as redraws their roles. Retailers retain the heavy lifting – logistics, service, last‑mile trust. Brands regain the early‑funnel context they lost and, for the first time, can inject richer first‑party data into a conversation that feels personal. If chat interfaces keep expanding their memory and intent modeling, they start to resemble an externalized CDP: Search engines still predict what we want, marketplaces still know what we buy, but the assistant begins to understand why. For marketers the lesson is simple: make your data and services addressable, semantically clean and ready for an agentic world. The conversation is moving upstream; meet it there or watch someone else define your customer’s next question.
