Reflections on MVP Summit 2026

I have had the privilege of being awarded the Microsoft “Most Valuable Professional” Award by Microsoft every year since 2009. One of the primary benefits of this award is the opportunity to visit the Microsoft campus in Redmon, WA, each year, and speak directly with the product teams who build the Microsoft products we use every day. I just returned from the 2026 MVP Summit and here are some of my thoughts and observations.

🔨We’re All Builders Now

As I was waiting in the TSA line on my way to the conference, I was reading a news article on my phone about how, in Silicon Valley, “everyone is a builder.”  I had not heard that term used before specifically, but I heard it used repeatedly by product managers at Microsoft. And it makes sense. You no longer are a “designer”, “coder”, etc., because these disciplines are morphing into a unified role, where you prototype the UX and ask AI to build it for you. (I suppose this follows on the heels of those who create things no longer being “crafters” but “makers”.)

🤖 Humans Calling Agents Calling Agents

Every product team is grappling with what to do with, and refer to, and track, “unattended” or “autonomous” (which certainly sounds more ominous) agents. To a certain extent, the problem is not new, in the sense that we’ve had code running behind the scenes since the beginning. If a user opens a file, we track the date it was accessed, but we don’t record when the search indexer accessed that file. In the past, this work was done at the code layer itself, or if done as a user principle, it happened as a service account. The difference is that agents are now working, asynchronously, on behalf of, a user. So does that “count” as being accessed by a user or not? What does that mean in terms of security, control, and reporting? Some teams wanted to call out human interaction as opposed to agent interaction, while others argue that they should be treated the same.

👶🏻👴🏽Innovation vs. Experience

One of the last sessions I attended was run by two fairly young product managers. They asked to “blue sky” prototype and dream how we wanted a brand new product experience to work. I found myself instantly getting in the weeds mentally, wondering why they didn’t consider piggy-backing their product onto Product A or Product B that already exist. All I could see were the headaches and confusion of trying to “recreate the wheel.” But then it dawned on me, that is exactly why innovation stifles. It’s hard for us “seasoned” professionals to get out of our own way and not think inside the boxes we’re used to. And yet, I also know I have lived through many product lifecycles and understand many of the challenges we have faced repeatedly. But you just can’t have both at the same time: “blank slate” thinking, and lived experience. It made me realize the fundamental pressure Microsoft is facing as one of the longest-established software companies who is feeling pressure to innovate.

💗We’re All Humans After All

I was frankly a little disappointed at how information work was being reduced to data, simply inputs and outputs. For a while, I felt like Microsoft was really trying to focus on the people the software is supposed to serve. This was especially so during and after COVID, when, for example, Viva was providing tools to prevent employee burnout, and help employers track employee work-life balance based on how often they checked e-mails after hours, etc. Compare that to this week, where the conversations seemed incredibly reductive, as if all work could be carried out by autonomous agents. In one session in particular, a project manager raised her hand and said, “I have been on projects that were late and over budget, but the customer was happy.” In other words, on paper it wasn’t a success, but in real life, it actually was OK because, because the project manager managed expectations and everyone was happy. In other words, tracking tasks and statuses and budget and cost are well and good, but at the end of the day, managing people is what makes a project successful.

In another session, a product manager talked about how we could offload his thoughts verbally into an AI agent that could note-take and brainstorm for him and called it his “second brain.” But contrast this with a conversation with a colleague of mine who said he’s in so many daily meetings that he can’t even remember the conversations he had with the same people a week ago. Is it healthy that we let AI log every interaction, when as humans, we can no longer even remember what we said because we are so overworked and mentally taxed? I believe the most successful AI will assist us to be better humans, not reduce us and our activity to data points.

One product manager even said within their team, they joked that they would just be writing agents that would work on behalf of one person, that would just talk to another agent working on behalf of another person, and no people would actually talk. Much has been said of this Dead Internet Theory (most notably by Sam Altman himself), where agents are both writing and reading content, but it still begs the question of, what are we actually trying to do here? Are we missing the forest through the trees?