Nicholas Salzman

I've spent the past decade leading product development teams through the kind of transformation that looks obvious in retrospect and impossible in the moment.

Bespoke Agile started in 2016 as a set of observations about why methodology adoption fails. Not the surface-level failures — missed standups, ignored retrospectives, Jira boards nobody updates — but the structural ones. Teams adopt a methodology designed for a different shape of team, then blame the methodology when it doesn't fit.

The framework that emerged from those observations was built around a simple premise: process should adapt to teams, not the other way around. "Bespoke" isn't a brand name — it's the operating principle. Every team's process should be tailored to their actual capabilities, constraints, and goals.

Then AI Changed the Question

When large language models moved from research curiosity to operational tool, the adaptive framework I'd been building turned out to be exactly the scaffolding needed. Not because I predicted AI — I didn't — but because the underlying problem is identical: teams need a process that accommodates a new class of capability without losing coherence.

The difference with AI is dynamic bottlenecks. In traditional development, bottlenecks are relatively stable — they move slowly as teams grow or shrink. With AI agents in the pipeline, bottlenecks shift constantly. An agent that generates code at machine speed creates a review bottleneck. An agent that drafts specs creates a coherence-checking bottleneck. The process has to adapt in real time, or it breaks.

That insight — that coherence, not velocity, is the metric that matters — is what drove BespokeTracker from concept to product suite.

What I Do Now

I build tools and advise teams navigating the transition to agentic product development. The article series lays out the argument. Comply, Coherence, and Codesign are the implementations. Consulting is the hands-on work of helping teams through the transition.

The common thread: helping people work with AI instead of being displaced by it. The technology is moving fast. The organizations adopting it don't have to move recklessly.