Introduction: AI will widen the gap between PMs, not steal their jobs
Generative AI has become commonplace, and the work of PMs (Product Managers) has changed dramatically. Specifications, requirements refinement, research, UI copy drafts, and backlog shaping—such tasks can be done faster by AI. But to that extent, the quality of decisions about what to build, why to build, and how to win will be more important than ever for product success.
This article summarizes the 10 essential skills for thriving PMs in the AI era in a friendly, approachable way that you can try starting tomorrow.
1. Problem Framing: Decide what to solve before feeding it to AI
The more you use AI, the more important the way you frame the problem becomes. Ambiguous questions yield ambiguous answers. A PM's value shows up in the initial way you frame the problem.
- Bad example: I want to reduce churn. Give me some initiative ideas.
- Good example: The Time to Value (TTV) for new users is long. Where in onboarding do users drop off? I want to split into three hypotheses and create a validation plan.
The trick is to pass AI a set of background, constraints, success metrics, and target users.
2. AI Literacy (Assessing capabilities and limitations): Manage expectations
PMs should at minimum understand that LLMs generate plausible text but do not guarantee truth. Hallucinations and biases from training data are realities in practice.
Additionally, multimodal AI handling text, images, audio, and video has become standard, making configurations like Retrieval-Augmented Generation (RAG) and agents increasingly common. Even if you can't implement everything, being able to explain the concepts conceptually greatly raises the resolution of conversations.
3. Data Thinking (Metrics design and measurement): AI products won’t grow unless you can measure them
Because AI outputs can vary, defining what constitutes “good” is harder than before. Therefore, PMs must own metrics (KPIs) and evaluation design.
- Product KPIs: Retention, churn, NPS, TTV, feature usage, etc.
- AI quality metrics: Accuracy, usefulness, reproducibility, toxicity, cost ($ / request), latency (response time)
- Operational metrics: Failure rate, escalation rate, human review ratio
Practically, link analytics platforms (Amplitude / Mixpanel, BigQuery, Looker, PostHog) with evaluation logs (prompt, context, output, user reaction) to sustain a continuous improvement loop.
4. Accelerating Hypothesis Testing (Experimentation): Use AI as a sounding board to speed up validation
In the AI era, PMs tend to win by the speed of validation rather than sheer idea volume. AI can accelerate the upstream work of research summaries, competitive comparisons, user interview questions, and PRD drafts.
The recommended flow is:
- Use AI to broaden the set of hypothesis candidates (10–20 ideas)
- PM narrows to three promising paths (strategy, customer value, feasibility)
- Verify with minimal experiments (fake door, prototype, A/B)
- Use logs and user feedback to decide the next move
5. Deepening User Understanding (JTBD/UX): The more AI exists, the more important human frictions become
Even in an era where AI can answer anything, users still face practical constraints like being bothered, anxious, embarrassed, or忙しい. PMs must capture these realities.
A framework that pairs well here is JTBD (Jobs To Be Done). Articulating what the user wants to achieve with the feature helps to keep AI functions from drifting.
Example: The value of an AI for meeting minutes is not just transcription; it speeds decision-making and ensures you don’t miss what needs to be done.
6. Workflow Design Beyond Prompts: Value lies in the process
Writing good prompts is convenient but for PMs it's only the entry. Real difference comes from how you embed AI into the user’s workflow.
- What is the input (text, PDFs, CRM data, audio)
- What is the output (drafts, candidates, decision materials, final deliverables)
- Where do humans verify and approve (Human-in-the-loop)
- How to revert on failure (fallback, retry flows)
For example, in sales email generation, designing the sequence from generation → internal tone check → legal-NG word check → CRM entry → send creates an experience where the workflow advances the business, not just produce text.
7. Tech Communication: Have the same map as engineers
AI features require more technical choices that influence product experience. PMs should keep these conversational points:
- RAG: internal document search + generation to increase accuracy
- Evaluation: offline evaluation (test sets) and online evaluation (real usage)
- Cost: model pricing, token reduction, caching, batching
- Latency: perceived speed, streaming, progressive rendering
- Permissions: handling confidential data, access control, audit logs
Even if you can’t write code, being able to express how a choice affects the UX and KPIs makes a PM strong.
8. Guardrail Design (Safety, Legal, Ethics): AI can scale accidents too
AI is convenient, but mistakes can scale as well. PMs should prepare the mechanisms to prevent mishaps in advance, which boosts trust from the development team.
- Privacy: handling personal information, training use, retention period
- Copyright: rights to generated content, quotation, and reuse
- Safety: inappropriate outputs, discrimination, prompt-injection countermeasures
- Transparency: explicit use of AI, justification (referencing sources when possible)
Especially in B2B, security requirements like SOC 2 and ISO27001 are common, so lock in with legal and security early.
9. Stakeholder Alignment (Decision-Making Style): AI opinions tend to split
AI features can look feasible, which can raise expectations. Sales may think everything can be automated, support fears misanswers, legal worries risk. In such moments, PMs are needed—not by stubbornness, but by a decision-making framework.
- What to optimize (growth? quality? cost?)
- Align assumptions (target users, usage scenarios, acceptable error)
- Assemble decision inputs (logs, PoC results, competitors, risk list)
- Decide how to decide (DRI, RACI, deadlines)
The PM who can design the decision-making process is especially valued in the AI era.
10. Storytelling (Translating Value): Describe AI value as an experience
Finally, a bit down-to-earth but super important. Explaining AI as “models” or “accuracy” tends to be dull. What executives and users care about is: how much easier will it be?
Strong PMs translate value like this:
- Feature explanation → before/after (work time from 30 minutes to 5 minutes)
- Accuracy → impact on operations (first-pass resolve rate +15%)
- Technical elements → reassurance (cite sources, audit logs, human verification)
This translation works for internal communications as well as outward communications (PR/sales).
Practical checklist for tomorrow
- Problem framing: Are you passing AI background, constraints, and success metrics together?
- Evaluation: Are you tracking AI quality metrics (accuracy, cost, latency)?
- Logs: Are prompts, outputs, and user reactions linked?
- Workflow: Have you designed the end-to-end flow (input → approval → storage)?
- Safety: Have you set prohibitions, fallbacks, and explicit AI usage disclosures?
Conclusion: PMs in the AI era become like editors-in-chief
As AI can mass-produce text and drafts, PMs become responsible for deciding what to adopt, what to discard, and how to connect things—essentially an editor-in-chief. They gather outputs, validate them, shape a tight storyline, and push the team forward.
You don’t need to perfect all ten at once. Start by building the loop: problem framing → metrics → evaluation → improvement. Even taking one step at a time will bring you closer to becoming a stronger PM.
