KM4BI KM4BI

Cheap software made your PM job harder, not easier. Here’s the new job.

As AI makes prototypes cheap and abundant, product management shifts from rationing engineering to governing what software should matter.

Nate B. Jones argues that AI is not merely turning product managers into faster prototypers. It is moving the bottleneck. When first versions become cheap, the hard work becomes deciding what should exist, what should be supported, and what should be deleted.

Prototyping is becoming table stakes

Tools such as Lovable, Claude Code, Codex, and low-code platforms mean that working artifacts can appear anywhere in the organization. A PM may no longer be the first person to turn an idea into software. A support rep, operations team, or business function may already have a dashboard, automation, workflow, lightweight app, or agent running before product is involved.

Software abundance needs classification

Jones uses Microsoft’s internal Power Platform footprint as an example of what this looks like at scale: more than a million internal assets across apps, automations, chatbots, and agent-like environments. The product question changes from “should engineering build this?” to “what class of software is this, and should the business rely on it?”

A production class ladder

The proposed answer is a ladder. A personal tool can remain scrappy. A team beta needs an owner, scope, touched systems, and a failure plan. A supported internal product needs access control, monitoring, documentation, support, auditability, and change management. A customer-facing feature adds the normal product bar plus AI-specific evaluation and governance where needed.

The PM job becomes more technical

This makes product management more strategic, but also more technical. Product decisions now depend on model behavior, agent loops, data access, retrieval, evaluations, latency, cost, permissions, reliability, and trust. PMs do not all need to become full-time engineers, but they do need to understand the systems well enough to make sound product judgments.

The takeaway

The post-prototype PM enables broad experimentation while creating an intentional promotion path for the work the company will depend on. The goal is neither central control nor chaos. It is judgment: identify real demand, separate useful signals from local convenience, and decide what deserves to become a real customer or business promise.

Source