Microsoft AI Culture Shift for Developers

Amazing work coming out of the Siemens, Microsoft and Slalom teams. Paving ways for how AI can revolutionize the Product Development Lifecycle.

Transformation doesn’t happen through technology alone. Together, Siemens, Slalom, and Microsoft built a culture of learning and innovation—equipping 30,000 developers with AI tools that are changing how work gets done: https://slalom.ws/3PcVIb4

Everyone says the framework decides your AI success. It doesn't. Your execution pipeline does. The framework wars are over because most modern stacks are good enough. The real bottleneck is messy data, weak evaluation, slow iteration, and poor deployment decisions. Winning teams follow one framework: collect clean data, test small, measure everything, automate feedback, then scale. That shift cuts failed AI experiments by over 40% and reduces deployment time by weeks instead of months. The next competitive edge won't be choosing another framework. It will be building systems that learn faster than competitors. #ArtificialIntelligence #SoftwareEngineering #MLOps #TechLeadership

Every enterprise scaling past one AI agent hits the same wall: reaching the systems it needs to touch. There are four common ways teams try to solve it today and a fifth that's fast becoming the standard. Swipe through, and see slide 8 for the part most teams skip entirely.

Everyone has the model. Almost no one has the context. That's why so many enterprise AI pilots die in production — and how Microsoft just fixed it. 👇

74% of companies still can’t show measurable AI ROI. In operations, that often shows up as time spent chasing updates, maintaining documentation, and managing repetitive reporting instead of focusing on higher-value work. Microsoft Copilot helps bring structure to that work by turning information from across systems into clear summaries and usable documentation—reducing manual effort and keeping workflows moving. Our craft is helping teams unlock the full potential of Copilot by focusing on the work that matters most. Watch our short AMA video with President + CEO Harry Zarek to see how organizations are identifying the right Copilot opportunities before they deploy. https://hubs.li/Q04mJ6kJ0 #MicrosoftCopilot #AIAtWork #OperationalExcellence #FutureOfWork #ServiceisOurCraft

As AI models evolve, GPT-5.6 is now powering core productivity tools like Microsoft Copilot 365, challenging us to rethink the balance between automation and human expertise. How are you navigating this shift in your organization?

As AI models evolve, GPT-5.6 is now powering core productivity tools like Microsoft Copilot 365, challenging us to rethink the balance between automation and human expertise. How are you navigating this shift in your organization?

Two things happened this week: 1. Anthropic made claude sonnet 5 the default for every FREE user. a model performing close to their flagship. cost to you: zero. 2. Microsoft raised M365 enterprise prices. official reason: "added AI capabilities." so let me get this straight. raw intelligence is getting cheaper every quarter. the software wrapped around that intelligence is getting more expensive. We're basically paying more for the box the free thing comes in. and honestly? it's a preview of the next 5 years. the model is not the moat anymore. everyone will have frontier AI in their pocket for free. the money moves to whoever packages it into something your team actually uses without a 3-week onboarding call. My prediction: by 2028 "we have AI features" will sound exactly like "we have a website" did in 2010. congrats, so does everyone. The interesting question isn't "does your product have AI" - it's "what does your product do that's still worth paying for when the AI is free?"

Most AI infrastructure founders should not copy the AI application playbook. I wrote this piece for technical founders building AI infrastructure, API/data platforms, and deep tech data platforms who need pricing, GTM, and unit economics to match the business they are actually building.