Diffuse Disciples→Focused Outcomes

How the real art of business is making demands coherent (and critical)

Product managers in modern software development face a peculiar challenge: they must master the art of saying no.  Mark Vogelgesang, currently a product manager at Salesforce and formerly a public servant, articulates this problem with precision. 

His domain encompasses discrete feature sets—from design systems to integration platforms—where the scope matters less than the focus. The real difficulty emerges when requirements flow in from every organizational level, from intern to C-suite, each voice demanding attention. The solution, he argues, lies in organizational alignment. Without it, refusal can become arbitrary, never a sign of a functional workplace. With alignment (read: a shared understanding of what everyone should be doing, from organizational top to bottom) secured, prioritization clarifies itself. 

But of course neither extreme is often visited. The real challenge lies in the everyday middle ground, where demands arrive from senior leadership but lack the supporting infrastructure below, or from well-meaning and insightful colleagues, whose ideas are often good but lack support from both above and below. In these situations, the product manager must talk trade-offs, getting the colleague on board with placement of their idea in the backlog, or work with management to re-shuffle priorities rather than accepting an impossible accumulation of high-priority work that serves no one.

The rise of artificial intelligence has complicated this already difficult task. This is because the barrier to using AI sits considerably lower than the barrier to understanding it. This asymmetry creates a serious problem: people can deploy AI tools without grasping their limitations or interrogating their outputs critically. 

He speaks from experience about code editors equipped with Model Context Protocol servers that can construct requirements, compile code, format commits, and iterate through testing cycles with remarkable objectivity. Yet this capability demands something often overlooked—critical evaluation. The machine may run tests that either pass or fail, but someone must still ask whether the right tests were written in the first place.
On the other end of the technological spectrum, the federal procurement space offers a similarly telling example: contracting officers, unfamiliar with technical nuances, might disqualify vendors for using synonyms rather than exact terminology. AI could bridge such gaps, but only if wielded by those capable of questioning its recommendations rather than accepting them wholesale.

This capacity for critical distance finds its roots in an unexpected place: breadth of experience. Mark's current reading—David Epstein's "Range"—argues that diversified backgrounds produce superior outcomes, pushing back against the cult of early specialization. Roger Federer didn't touch a tennis racket seriously until age 14, yet became the sport's greatest player. Mark's own liberal arts education fostered constant interdisciplinary activity, training him to pull ideas from disparate fields and synthesize novel solutions. This wide-ranging foundation matters more than ever in an AI-saturated environment.

The balance required of modern product managers—weighing competing demands, navigating organizational politics, maintaining extensible architectures—grows more precarious as AI amplifies both capability and risk. Mark's liberal arts background positions him ideally to meet this moment: trained to question, to synthesize, to see connections across domains. In an era when machines can write code but not yet wisdom, the critical stance he brings becomes not merely valuable but essential.

What we’re into this week

Scott

Mark’s insights reminded me of this great article on the Strangler Vine approach to digital product evolution. It requires that the team really understand the current state of the users and the product in order to build something better that’s also not disruptive; you’re not going to get that without a deft product manager, and definitely not with—to use Mark’s word—a “greenfield” approach. 

Ana

I think there are so many ways to approach an issue, and it does so much depend on context. For example, Mark talked about range using tennis player Roger Federer as an example, and he’s cited a lot in these conversations. But you can’t have a conversation about excellence without talking about Serena Williams, whose father famously did push her and her sister from a very young age. And then there’s Misty Copeland, who didn’t start dancing until she was a teenager. The TL/DR here that there’s just not one effective approach; it depends on context. 

Aaron

You all are both so insightful here—I’ve just been at my desk using AI to produce slides. 🙃 

CX Research—Documented

Wondering about prioritization? One of the best ways to find out what your internal and external customers want is to simply ask them. Find out how to do just that using the HCD Discovery Guide, the only real-world tested guide to meaningful, sustainable customer research and insights. 

Buy the book at Barnes and Noble, Amazon, or at Ishmael Interactive. 

You don't need a full UX or CX team to get to know your customers: you need the HCD Discovery Guide.

Credits

Guest: Mark Vogelgesang
Host: Aaron Meyers
Producer: Ana Monroe
Text: A Adams
Artwork: Diagram of a Lunette [verso] 1890/1897. Charles Sprague Pearce.
Why this artwork?: Planning can be beautiful, and should be.

Show notes:

Mark Vogelgesang, product manager at Salesforce, discusses the critical art of saying no in software development and how organizational alignment determines what gets built. He explores how AI tools like code editors with Model Context Protocol are transforming development workflows—but warns that the ease of using AI far outpaces most people's ability to critically evaluate its outputs. Drawing on David Epstein's "Range," Mark explains why his liberal arts background and interdisciplinary thinking give him an edge in an industry increasingly shaped by artificial intelligence.