👋 Good morning! AI isn’t just learning tricks anymore, it’s stepping into real-world roles. From multilingual agents that navigate cultural nuance to grocery assistants that plan your meals, and from high-speed language models to AI-generated chart-topping musicians, this week’s edition explores how intelligence is moving from theory into practice. Strap in, AI isn’t just observing, it’s participating.
🧙“Wonderful” is Building Culturally Aware AI Agents at Scale
An Amsterdam-headquartered AI startup that’s rethinking how companies engage with customers, not just in one language, but in many, and with cultural nuance built in. In November 2025, the company raised $100 million, pushing its valuation to $700 million.
What makes Wonderful special is its focus on multilingual, multi-modal AI agents that can handle customer interactions across email, voice, and chat, and importantly, do so in a way that adapts to local cultures. Rather than using generic chatbots, their agents understand not only what to say, but how to say it depending on the region.
That’s not just a cosmetic layer: as the demand from enterprises has surged.. After all, companies are increasingly global, and customer expectations change from market to market. Wonderful’s CEO, Bar Winker, points out that enterprises are already using their agents for real volume.
The $100M funding will primarily fuel two things: local hiring and further tech development. They’re doubling down on building out R&D teams that understand regional nuances, while also scaling up the core AI platform.
Early results are promising: Wonderful expects to hit $10 million in annual recurring revenue by the end of 2025, backed by fast adoption in Europe and the Middle East.
Why it matters: We often talk about AI in terms of “make it faster” or “make it smarter.” Wonderful is showing another dimension: “make it local.” Their model hints at a future where AI isn’t just a tool for automation, it’s a bridge between cultures. That could be a game-changer for enterprises operating across borders.
⚙️ In Focus: AI Tools & Products
Inception Labs announced a major upgrade to its diffusion-based language model, Mercury, with a new release that improves its performance across coding, instruction-following, math, and knowledge tasks.
What makes this version of Mercury exciting is its diffusion architecture: instead of generating text token-by-token (like traditional LLMs), it starts from a “noisy sketch” and refines entire blocks of text in parallel. This enables up to 10× faster inference compared to many speed-optimized autoregressive models.
It’s available via API (and compatible with the OpenAI API spec), so developers can plug it into existing workflows.
On Azure, Mercury is production-ready through Azure AI Foundry, giving enterprises access to secure, scalable infrastructure.
Key takeaway: Mercury demonstrates that diffusion-based LLMs are no longer just a research curiosity — they’re maturing into high-performance, cost-efficient tools ready for practical applications like real-time code assistants, responsive chat agents, and systems that require low latency.
🤖AI in Action: Instacart’s Cart Assistant Transforms Grocery Shopping
Instacart has launched a new suite of enterprise AI tools, including a conversational agent called Cart Assistant, purpose-built for grocery shopping.
Here’s how it works: Cart Assistant is deeply integrated into both Instacart-powered e-commerce sites (via Storefront Pro) and in-store smart carts (Caper Carts). Rather than being a basic chatbot, it acts like a personalized shopping companion: it helps customers with meal planning, suggests products based on their preferences and history, keeps an eye on budgets, and even supports nutritional goals.
On the retailer side, this is part of a bigger AI strategy. Instacart’s launch includes five core pillars:
Agentic Commerce — building next-gen shopping agents with generative AI.
In-Store Intelligence — using computer vision in stores; Caper Carts get live shelf visibility, heat maps, and inventory tracking.
Catalog Intelligence — the “Catalog Engine” enriches product data (nutrition, ingredients, etc.) using vision models.
Agentic Analytics — giving retailers tools to query their data using AI agents, turning complex data into quick insights.
Why this matters:
For shoppers: AI feels more like a helpful assistant than just a recommendation engine, it adapts to how you shop, what you eat, and how much you want to spend.
For retailers: the technology gives real-time store insights, improves catalog data quality, and provides a way to turn data into actionable decisions quickly.
For operations: combining AI with smart carts and computer vision opens up new ways to reduce friction in-store and make shopping more efficient.
Takeaway: Instacart is not just using AI to suggest things, it’s embedding it into the entire shopping journey, from planning meals to navigating the store. This is a concrete, high-impact example of AI working at the intersection of physical retail and digital commerce, not automating for automation’s sake, but making shopping smarter, more personal, and more efficient.
💡Quick Hits and Numbers
According to KPMG’s Q3 2025 survey, 42% of organizations have deployed some form of AI agent (up from 11% two quarters earlier).
Workers with AI-related skills (like prompt engineering) command a 56% wage premium.
The AI‑generated country artist Breaking Rust hit No. 1 on the Billboard Country Digital Song Sales chart with the single “Walk My Walk.” The entire persona, voice, and music were created by AI.
🧩Closing thought
AI is no longer a background helper; it’s a collaborator, a guide, and sometimes an unpredictable innovator. Whether bridging cultures, optimizing shopping journeys, or composing music, its growing presence forces us to rethink control, trust, and creativity, not just what AI can do, but what it should do. The future won’t replace humans; it will replace human limits.