CAREERSOPERATING_MODELS4 October 20247 min readLee Leckenby // System Builder

Why Product Managers will thrive with AI

AI does not replace Product Managers. It rewards the ones who think in systems, write clearly, and measure outcomes. Here’s how to use it without turning your workflow into fluff.

// FOCUS

Using AI to increase product management leverage.

// AUDIENCE

Product Managers.

// FORMAT

Perspective.

The rise of the machines

The loudest AI debate is about replacement.

The practical story is about leverage.

Product Managers sit in the middle of messy systems: users, data, engineering constraints, business goals, incentives. AI is strongest when the work is messy and ambiguous. That’s why PMs are well-positioned.

Not because we are “visionaries”.
Because we are translators with receipts.

PM work is already model-friendly

A useful way to think about it:

AI is good at drafting and pattern-matching.
PMs are good at choosing and correcting.

That pairing is the point.

If you ask AI for “ideas”, it will give you a pile of ideas.
If you give it context, constraints, and a definition of success, it starts producing work you can actually use.

That is the same move we make with stakeholders.

We already think in systems

PMs spend a lot of time doing three things:

  • spotting patterns across noisy inputs

  • forming hypotheses

  • turning chaos into structure

That maps well to how AI helps:

  • summarise and cluster information

  • propose options and trade-offs

  • generate structured first drafts

The difference is speed.

AI will produce a “good enough” structure in seconds.
Your job is to decide whether it matches reality.

We already prompt, we just call it “writing”

A ticket is a prompt.
A PRD is a prompt.
A stakeholder brief is a prompt.

Same mechanics:

  • define the goal

  • provide context

  • add constraints

  • specify output format

  • iterate until it is usable

If you want a simple starting prompt that behaves:

You are a senior Product Manager for a marketplace.
Goal: increase buyer trust without slowing checkout.
Constraints: no new compliance approvals, 2-week build, measurable impact.
Output: 3 options, each with risks, success metrics, and rollout plan.

That’s not “prompt engineering”.
That’s clear product thinking.

AI is best at the boring parts

The fastest wins are not the flashy ones.

They are the tasks that drain energy and time:

  • synthesis from calls, surveys, tickets

  • writing first drafts of specs and one-pagers

  • generating edge cases you forgot

  • turning messy notes into stakeholder updates

  • drafting experiment plans and instrumentation checklists

This is the structural insight:

AI makes you faster at turning input into artefacts.
It does not make you better at deciding what matters.

So use it to clear the undergrowth.
Keep your attention for the decisions.

PMs are translators, AI needs translators

Good AI outputs usually require:

  • the right level of detail

  • the right trade-offs

  • the right vocabulary for the audience

That is literally PM work.

You can translate in both directions:

  • engineering notes → customer-facing release copy

  • user feedback → structured themes and hypotheses

  • roadmap ideas → measurable bets with risks and dependencies

  • a feature brief → acceptance criteria and test cases

Think of AI as a junior stakeholder who never sleeps.
Helpful. Fast. Needs direction.

We are measured by outcomes, so we can keep AI honest

The easiest way to waste time with AI is to optimise for outputs.

More docs. More slides. More “ideas”.

PMs win when outcomes move.

So keep the loop tight:

  • define the metric before you ask for help

  • ask AI for options that connect to the metric

  • validate with real data and real users

  • ship the smallest useful thing

  • measure, learn, iterate

AI is not the strategy.
It is the accelerator for your learning loop.

Quick wins you can try this week

  • Research synthesis: paste anonymised notes and ask for themes, contradictions, and “what we still don’t know”.

  • Roadmap critique: ask it to stress-test your roadmap against a persona, a constraint, and a business goal.

  • Stakeholder updates: draft a one-page update with decisions, risks, asks, and next steps.

  • Prioritisation assist: have it apply RICE or a value vs complexity sort, then challenge the assumptions.

  • Edge-case hunt: ask “what breaks here?” and “what could harm users?” before you ship.

Keep it small. Keep it grounded. Keep it measurable.

Final thoughts

AI does not replace Product Managers.

It changes the skill gradient.

The PMs who thrive will not be the ones who generate the most content.
They will be the ones who:

  • constrain the problem cleanly

  • iterate fast

  • validate with discipline

  • tie work to outcomes

One implication for builders: if you want AI to make you better, build a repeatable loop around it.

Prompt. Draft. Verify. Ship. Measure.