ai and careers
In the short term, Gen AI PM roles will pay more. In the long term, no. The premium disappears once everyone has the skill. The question is what you build underneath it.
Let me tell you what nobody in the PM training industry will say plainly: AI is going to eliminate a significant portion of what junior and mid-level PMs spend their time on. Not eventually. Now. It is already happening at the companies paying attention.
The question is not “will AI affect PM careers?” That debate is over. The question is: which parts of your current role are getting automated, which parts become more valuable, and what you should be building in your career right now — before the gap between PMs who understand this and those who do not becomes unbridgeable.
This is not a comfort piece. It is an assessment.
What is actually getting automated
The 2024 job market for PMs in India looked strange. Headcount at mid-size tech companies was flat or shrinking, but the remaining PMs were shipping faster than ever. The explanation is not magic. It is that a large category of PM work — the scaffolding work, the prep work, the documentation work — is now done by AI in a fraction of the time.
Specifically:
Requirement drafts and PRDs. A PM who used to spend two days writing a requirements document can now produce a first draft in two hours using Claude or GPT-4o, then spend the rest of the time in the room validating and refining it. This does not mean the PRD skill is worthless. It means the emphasis has shifted from writing to judgment about what goes in the document.
Competitive research synthesis. Spending a week reading competitor changelogs, App Store reviews, and product blogs used to be standard PM work. AI summarises this in an afternoon. The job is now knowing what to look for and what it means, not the reading itself.
User interview analysis. AI can cluster themes from twenty user interviews, surface recurring pain points, and generate a structured synthesis document. The PM’s job shifts to designing the interview guide, running the conversation, and making the judgment call on which insights actually matter for the roadmap.
Documentation and release notes. Sprint summaries, stakeholder updates, release notes — these are being handled partially or fully by AI at forward-thinking product teams. The ones who haven’t automated this yet are falling behind.
Estimation and sizing. Rough effort estimates, market sizing with public data, RICE scoring on feature lists — AI produces these in minutes. They need human sanity-checking, but the grunt work is gone.
Here is the honest implication: if you are a mid-level PM whose value comes primarily from being organised, thorough, and fast at documentation, your advantage over AI is diminishing. That is not a prediction. That is this year.
Q3 planning. A product team at a Bangalore B2B SaaS company, 80-person startup.
CPO: “The board is asking why we have five PMs when Competitor X ships just as fast with two.”
PM Lead: “Because we do discovery, documentation, coordination, analytics, stakeholder management, roadmap —”
CPO: “Which of those can an AI do 80% of, given the right prompt and context?”
PM Lead: “...the documentation. The analytics summaries. First drafts.”
CPO: “So what are you doing with the time you save?”
Nobody had an answer. That silence is the real problem.
The question is not whether AI takes your job. It is whether you are using AI to free up time for higher-impact work — or just doing the same work slightly faster.
What becomes more valuable
There is a category of PM work that AI makes more important, not less. These are the things AI cannot do — not because of technical limits, but because they require standing in a room, reading the room, and being responsible for a decision.
Taste and judgment. AI generates options. Lots of them, on demand, at high quality. The bottleneck is no longer generating the options — it is picking the right one. This requires product taste: the accumulated sense of what works for your users, what your brand stands for, what your engineering team can actually build well. Taste is not learnable from a framework. It is built over years of shipping things and watching what happens.
Customer intuition. The PM who has spent three years talking to small business owners in Tier 2 Indian cities knows something that no AI has been trained on: the specific texture of that user’s frustration, trust, and constraint. AI can analyse interview transcripts. It cannot replace the intuition built from being in the room when a user’s face changes when they see a prototype that finally gets it right.
Cross-functional alignment. Shipping a product requires getting engineering, design, sales, and leadership aligned on a direction — often when those groups have competing incentives and different information. This is fundamentally a human coordination problem. AI can draft the communication. It cannot read the political dynamics, manage the ego of the VP who didn’t get their priority funded, or build the trust that makes alignment possible.
Ethical and strategic call-making. Who benefits from this product? Who is excluded? What happens if this feature gets misused? These questions sit at the intersection of values, strategy, and context. AI can surface considerations. It cannot make the call.
Navigating ambiguity. The situations where a PM is most valuable are precisely the ones where the data is incomplete, the direction is unclear, and someone needs to make a bet. AI performs best on well-defined tasks. The ill-defined, high-stakes moment — a competitive threat, a major pivot, a team crisis — is where human judgment matters most.
The India context
This matters differently in India than in the US, and anyone telling you otherwise is not paying attention to the local market.
The Indian PM market has a structural excess of PMs who were hired for coordination and documentation roles during the 2021-2022 startup funding boom. Many of those roles existed because companies were moving fast and needed people to manage communication across large, geographically distributed teams. AI is compressing that need.
At the same time, there is a genuine shortage of PMs who can do the harder work: discovery in markets where users don’t speak English, strategy for India-specific constraints (payment rails, infrastructure, regulatory complexity, price sensitivity), and cross-cultural alignment between Indian engineering teams and US or UK stakeholders.
Gen AI PM roles — specifically the roles where PMs own AI feature development, AI content pipelines, or AI-powered user experiences — are paying a premium in 2024 and 2025. Talvinder’s read from coaching cohorts at Pragmatic Leaders: that premium lasts 18-24 months before it normalises, as Gen AI becomes table stakes rather than a specialisation. The window is open, but it is not infinite.
The PMs who will be well-positioned in 2027 are not the ones who learned to prompt ChatGPT. They are the ones who used the time AI freed up to develop judgment, intuition, and strategic depth — and who can demonstrate that depth with a track record of decisions that actually worked.
Take the last two weeks of your working time. Break it into categories: writing and documentation, research and analysis, coordination and meetings, decision-making and strategy, customer and user time.
Now ask: which of these categories is AI already doing 50%+ of, at equivalent quality, in 20% of the time?
Whatever time you save there — where is it going? Is it going into the categories that AI cannot replace? Or is it going into doing the same document-heavy work slightly faster?
If your time-shift ratio is not moving toward judgment, customers, and strategy — that is the problem to solve. Not which AI tool to learn next.
What does not work
There are two failure modes I see consistently in how PMs are responding to AI.
The tool-collector approach. Learning every new AI tool as it launches, adding them to a LinkedIn Skills section, and assuming that familiarity with Perplexity, Notion AI, and Linear’s AI features constitutes an AI strategy for your career. It does not. Tools change. The underlying judgment about when to use AI, what to trust, and where human oversight is non-negotiable — that is the durable skill.
The wait-and-see approach. “AI won’t replace PMs because PMs do creative, relational work.” This is technically true and practically useless as a strategy. The PMs who are using AI aggressively are shipping more, learning faster, and building more impressive track records — regardless of whether AI “replaces” PMs as a category. The floor is moving up. Staying at the old floor is falling behind.
The PMs who are figuring this out are not worried about whether AI will take their job in the abstract. They are asking: “What would it mean for me to use AI so well that I can do the work of 1.5 PMs, and use the recovered time to build the judgment and relationships that compound over time?”
That is the right question.
What to do now
Three concrete moves, not motivational advice:
1. Automate your own documentation, this week. Build prompts for your most common documents — PRD first draft, sprint retrospective, stakeholder update, competitive summary. Use them consistently for 30 days. Measure the time saved. Then consciously redirect that time.
2. Increase your discovery velocity. Use the time you save to run more customer conversations — not AI-generated surveys, but actual conversations. The PM with 20 real customer conversations this quarter has an edge that no amount of AI-generated synthesis can replicate. India has a massive diversity of users across geographies, languages, and income levels. That ground truth is available to any PM willing to go get it.
3. Pick a strategic depth to develop. Judgment about pricing strategy for Indian SaaS. Intuition for Tier 2/3 user behaviour. Understanding of the regulatory environment for fintech or healthtech. AI can brief you on these topics. Deep expertise — the kind that lets you make the right call under pressure — requires sustained focus over 12-18 months. Pick one and go deep.
Your company's VP of Product pulls you aside. 'I've been playing with AI assistants for a week. Honestly, they can draft specs, summarise research, and write release notes. That's a lot of what junior PMs do. I need your honest assessment — how should we be thinking about headcount going forward?'
You have 30 seconds to frame your response. You have two junior PMs reporting to you. What do you say?
your path
The honest summary
AI is not going to make PMs obsolete. But it is going to make a significant portion of current PM work look like data-entry looked after spreadsheets arrived — still necessary in places, but not a skill that commands a premium.
What remains premium: the ability to identify the right problem when nobody agrees on it. The ability to get a room of people with competing interests moving in the same direction. The taste to know which of a hundred AI-generated options is actually right for this user, this moment, this constraint.
That is the job. AI is changing what you spend your time on to get there. It is not changing the destination.
The PMs who treat this as a tool question — which AI should I learn? — will be fine for a year or two. The PMs who treat it as an amplification question — how do I use AI to build the judgment and relationships that compound? — will be the ones setting the agenda in 2028.
Next steps:
- AI Fundamentals for PMs — how the underlying technology works
- AI Product Strategy — building products where AI is a core feature
- Deliberate Skill Building — how to develop judgment intentionally
- IC to Senior PM — the shift from execution to judgment
A PM at a mid-size Indian SaaS company asks their head of product: 'Should I invest time in learning SQL and data skills, or in learning how to prompt AI tools effectively?' They have 10 hours per month for learning.
The call: What do you tell them?
A PM at a mid-size Indian SaaS company asks their head of product: 'Should I invest time in learning SQL and data skills, or in learning how to prompt AI tools effectively?' They have 10 hours per month for learning.
The call: What do you tell them?