Blog

The next generation of AI product builders may come from product management, from sales, or from the space opening up between them. It can work, but only with real guardrails: review, clear roadmap ownership, and incentives that reward lasting value.

My prediction rests on three observations.

1. Safe harnesses won’t stop at product managers

Companies want PMs to build with AI: fewer handoffs, smaller teams, more ownership. LinkedIn’s Tomer Cohen, then its chief product officer, put it this way: “We’re going to teach them how to code, design, and PM at LinkedIn” (Business Insider).

To let non-engineers ship safely, engineers build a development harness: proven patterns, automated tests, security controls, reviews, and recovery. Once it works, why would access stop at PMs? Extending it to sales is my prediction, not LinkedIn’s announcement.

Imagine a rep hears, “We’d buy this if it supported our approval workflow.” Today that becomes a feature request. Increasingly, it could become a working prototype, validated with the prospect and sent through review.

2. Roles are converging, and some already straddle the gap

Solutions engineers and forward-deployed engineers already build for customers. What’s changing is who can. As AI handles more specialized work, the lines between spotting a need, building a solution, and selling it keep shrinking. The Wall Street Journal reports on small cross-functional pods of people and AI agents, an early sign of that shift (WSJ).

3. Pay needs to follow contribution

This one is my view, not a finding. If AI-enabled employees are asked to multiply the value they create, it’s fair for pay to grow with the results. Commission already works that way. Fortune reported Danish research finding that only 3%-7% of AI productivity gains reached workers as higher pay (Fortune). That supports the incentive question. Revenue-linked pay is my proposed answer, not the study’s conclusion.

Many PMs also earn bonuses or equity, so the gap isn’t absolute. But those pay out on delayed annual or multi-year cycles. AI now lets a builder show results in days linked to specific customer outcomes, and commission is the structure that pays at that speed.

The catch: protect the product

A buyer’s request doesn’t automatically belong in the product. Without guardrails, you get one-off builds, tech debt, and promises that other people have to keep. Four things have to be true:

• Engineering owns review and maintenance.
• Product owns the roadmap. A prototype is a proposal, not a commitment.
• Legal and security review anything built before a contract is signed.
• Incentives reward durable value. One idea is to tie part of the payout to renewals, not just the first close. Attribution needs an explicit rule, because one built feature can help win deals for years.

The future is a hybrid

Product people become AI-enabled sellers. Salespeople become AI-enabled builders. Both move toward someone who can discover demand, build a solution, sell it, and share in the value they create.

That is close to a founder. Companies that don’t offer real upside risk losing these people to companies of their own.

So when your best builder asks why their upside is capped, what’s your answer?

Photo by Scott Graham on Unsplash.