Hello, Visionary CTOs! 🌟

What if AI didn’t just help you work faster… but actually knew how you work? 🧠

That’s not sci-fi—it’s already happening.

This week, we’re spotlighting three AI breakthroughs that hint at something bigger than automation. Citi’s CTO is quietly transforming one of the world’s largest banks using frontline feedback and “AI accelerators.” Microsoft’s Copilot just became your digital concierge—with memory. And one developer may have unlocked the secret to teaching AI how to think like a senior engineer.

These aren’t experiments—they’re signals. And if you’re a CTO, builder, or strategist, it’s time to look closer.

📰 Upcoming in this issue

• Citi’s CTO Has a Plan—and 150,000 Employees Are Already In On It 🧠

• Microsoft’s Copilot Just Got a Whole Lot Smarter (and a Little Creepier) 🤖

• When AI Starts Thinking Like a Senior Developer 🧠 

📈 Trending news

• CTOs, Take Note: Deleting Features Might Be the Best Optimization You Make All Year

• When AI Starts Running the Backend: What Higher Ed’s Shift Means for Tech Leaders

• UV Is Rewriting Python’s Packaging Rules—And CTOs Should Pay Attention

Citi’s CTO Has a Plan—and 150,000 Employees Are Already In On It 🧠 watch the full 8-min video here

Video published: April 9, 2025

In How Citi’s CTO is rolling out new gen AI productivity tools to more employees across the globe, Fortune gives us a rare look into how David Griffiths, Citi’s Chief Technology Officer, is reshaping a 230,000-person bank with precision and purpose.

Griffiths isn’t just dropping tools like Citi Assist and Citi Stylus into inboxes—he’s measuring AI impact by calculating how much human effort the tech replaces. Every new feature, from enhanced Q&A to real-time editing, is born from frontline feedback. His “AI accelerators”—handpicked employees embedded across departments—are turning change resistance into innovation at scale.

Griffiths is also eyeing the future: agentic workflows that translate, summarize, and even build slides are in the lab now. And unlike other tech leaders, he’s unapologetically pro closed-source AI. For Griffiths, it’s all about trust, security, and business value—no hype needed.

Key Takeaways:

• 🧠 CTO David Griffiths tracks “AI capacity” by comparing human vs. AI task costs. A cold, clear ROI model.

• 🌍 He’s expanded AI tool access to 11 countries—soon, a majority of Citi’s 230K workforce. Precision scaling.

• 🔁 “AI accelerators” act as embedded tech influencers. Real people driving feedback loops and adoption.

• 🧩 Prefers closed-source, multi-modal AI for control and compliance. Griffiths is building for resilience, not just buzz.

Microsoft’s Copilot Just Got a Whole Lot Smarter (and a Little Creepier) 🤖 read the full 1,130-word article here

Article published: April 4, 2025

In Microsoft’s Copilot can now browse the web and perform actions for you, TechCrunch reports that the chatbot once known for drafting emails just got a serious upgrade—and it’s gunning for your browser, calendar, camera, and maybe your to-do list, too.

Now able to take real-world actions like booking travel or ordering flowers, Copilot acts more like a personal assistant than ever. Microsoft has added memory for personal preferences, live camera analysis on mobile, and even voice-driven podcast summaries. It can now organize research into smart "Pages," and even respond to questions mid-conversation in its synthetic podcast mode.

But don’t worry—Microsoft insists you remain in control. You can delete memories or opt out entirely. With Copilot now embedded across platforms, from Windows desktops to Android cameras, the line between helpful AI and omnipresent assistant just got thinner.

Key Takeaways:

• 🛒 Copilot can book, reserve, and buy from partner sites like Expedia, OpenTable, and 1-800-Flowers. It acts like a digital concierge.

• 👁️ New camera vision mode answers real-time questions about what your phone sees. Example: “What flower is this?”

• 🧠 The bot now remembers preferences like favorite foods and delivers tailored, proactive suggestions. You can opt out via a user dashboard.

• 📝 "Pages" and “Deep Research” features mimic ChatGPT Canvas and Gemini, turning scattered info into polished outputs. Organized chaos, solved.

When AI Starts Thinking Like a Senior Developer 🧠 read the full 2,032-word article here

Article published: April 7, 2025

In The day I taught AI to think like a Senior Developer, Namanyay Goel drops a quietly radical idea that should be lighting up every CTO’s dashboard: what if AI could understand a codebase instead of just autocomplete it?

This isn’t about flashy AI demos. It’s about architectural literacy—the difference between AI that mimics and AI that reasons. By introducing Prismatic Ranked Recursive Summarization (PRRS), Goel simulates the way a senior engineer sees the big picture: architecture, data flow, security—each through its own lens. For CTOs tasked with scaling dev teams, onboarding talent, or upgrading legacy systems, this is more than a developer hack. It's a new blueprint for context-aware AI systems that support engineering at scale.

Forget prompt engineering. This is AI with a worldview—and that worldview might just change the CTO playbook.

Key Takeaways:

• 🔍 PRRS (Prismatic Ranked Recursive Summarization) lets AI see codebases through multiple lenses like architecture, data flow, and security. It’s like giving your AI assistant senior dev intuition.

• 🧠 This turns code from flat files into a mental model—similar to how real developers think and work. AI starts understanding reuse, abstraction, and pattern consistency.

• 🚀 CTOs can use this for faster onboarding, cleaner architecture, and surfacing technical debt automatically. It’s a force multiplier for code quality and engineering velocity.

• ⚠️ As AI gets smarter, mid-tier coding roles risk being squeezed—CTOs must rethink talent strategy. AI won’t replace devs, but it will shift which skills are most valuable.

Why It Matters

The leaders in these stories aren’t chasing trends. They’re rethinking what it means to lead, build, and scale with AI—not around it.

Whether it’s redefining employee workflows at Citi, shifting the boundaries of assistant tech at Microsoft, or giving AI a developer’s intuition—these breakthroughs mark a turning point. One where the real competitive edge lies not in using AI, but in shaping how it's used.

The smartest organizations won’t wait for the dust to settle. They’ll start experimenting now—because the future of work is already being written in code, context, and a whole new kind of intelligence.

Rachel Miller
Editor-in-Chief
CTO Executive Insights