The Senior Product Manager, as Advertised

This year those opinions do not even agree with each other.

In March 2026, Lenny Rachitsky counted over 7,300 open PM roles at tech companies globally, 75% above the early-2023 low and up nearly 20% since the start of the year.

Two months later came a survey of 1,500+ product leaders behind the 2026 CPO Insights Report from Products That Count and Mighty Capital, announced in the authors' own press release under the headline that product managers will "disappear by 2030". The release talks about the boundaries between product, design, and engineering "collapsing" and a hybrid PM-engineering "Product Builder" role taking their place.

Then there is Productside, which reports in its State of AI for Product Management Report 2026, a survey of 250+ product professionals, that 4 in 5 PMs use AI every day and only 1 in 4 have a strategy to show for it.

So demand is up, the role may be gone in four years, and everyone uses AI without much of a plan. I didn't feel like adding one more opinion to that pile. What I wanted to know was simpler. When a company opens a Senior Product Manager position, what does it actually write down? What does it require, and what does it expect the person to do, at least on the day the posting goes live? That is something I could check.

If you are outside tech, a quick baseline: a product manager decides what a team builds and why, and a senior one does it with a bigger scope and less supervision. That is the job these postings describe.

The dataset

I analyzed 605 Senior Product Manager postings, one per company, collected between June and September 2026 from company career pages and a few job portals.

Before the numbers, two things about how I read the postings.

• I split each posting into sections (what you will do, what we need, nice to have) and matched them against dictionaries of duty, requirement, and tool categories. When I write that 45% of companies ask for something, it means 45% mention that category in the relevant section. I found clean sections in about 85% of postings; for the rest I matched the whole text, which can add false positives. The dictionaries also miss roughly a quarter of the bullet points, which adds false negatives. So read every share as a directional estimate, not an exact value.
• The companies are mostly US and European tech firms whose career pages I follow. The numbers describe them, not the market.

I could assign an industry to 572 of the companies. The chart covers the 22 industries with at least ten companies each, 521 companies in total.

Fintech, SaaS, and healthtech alone make up a quarter of the corpus. When a number below starts to look like "the market," remember that.

In short, the posting describes a coordination job and asks for proof that you have done it before. About a third of the companies list the same five duties, four in ten ask for AI experience in the required qualifications, five years is the standard, and 69% name no tool or technology at all. Everything below is the detail.

What a senior PM is required to have

I started with the requirements section, because that is the part of the posting written with the applicant in mind. The chart shows the share of companies that ask for each category.

A few things caught my eye.

• Experience comes first, skills later. Explicit years (78%), prior PM experience (73%), and a proven track record (65%) all rank above any specific skill. Most senior postings are really asking one question: have you done this before?
• The first technical item on the list is AI. Familiarity with AI, ML, or LLMs (56%) ranks above APIs and platforms (48%), data skills (47%), and technical fluency (34%). What "familiarity" means in practice I come back to below.
• Domain expertise, a broad category that covers any named industry, shows up in more than half of the postings (55%). Senior PM roles open to "any industry" are rarer than the generic career advice suggests.
• The things people like to argue about sit near the bottom. Only 27% of companies ask for a degree, 24% for an engineering background, 19% for Agile methods, and 3% for certifications.

What a senior PM is expected to do

The duties section describes the job the company thinks it is hiring for, and here the postings agree with each other much more.

The classic PM core is all there: cross-functional collaboration, roadmap, working with engineering and design, end-to-end ownership, strategy and vision, prioritization. If I had written this list ten years ago, the top would have looked the same.

What I didn't expect is how quickly building AI products or features became a stated duty. 35% of companies already list it, about as often as shipping fast or making data-informed decisions. Discovery and user research, the part of the job the product community writes about most, appears in 29%. Postings describe outputs rather than practice, so I would not read that as "nobody does discovery," but it is clearly not what companies lead with.

Mentoring appears in 12% of postings and people management in 1.5%. A senior posting rarely asks the person to grow other people.

A recurring core

Because the same categories kept coming up, I checked how often they come up together. I went through every combination of five duty categories and counted the companies that mention all five.

The most common combination is cross-functional collaboration, roadmap, working with engineering and design, product strategy and vision, and prioritization. 188 companies, 31% of the corpus, list all five. On the requirements side, the most common five-item set (explicit years, PM experience, collaboration, track record, and comfort with ambiguity) appears in 24.5% of the postings.

57 companies, about one in eleven, have both of those sets in the same posting. So the Senior PM posting has a recognizable core, and roughly a third of the companies publish some version of it.

Underneath that there is variation. I ran an exploratory clustering on the postings, using which categories each one contains as its features (k-means; two clusters scored best by silhouette, though the score itself is low at 0.07, and in bootstrap resamples the groups kept about 92-93% of their membership; five near-empty postings could not be placed). The split is weak: most postings share the same core and the groups only differ at the edges. So read them as two contrasting bundles of language in this sample, not two kinds of company:

• Process, alignment, and analysis (318 companies, 53%): stakeholder alignment, business cases, market analysis, Agile, a degree. Financial services, retail, industrial, e-commerce, and consulting lean heavily this way; fintech is split down the middle.
• Shipping and AI (282 companies, 47%): fast iteration, AI features, product sense, startup pace, a hands-on record of shipping. SaaS, devtools, AI companies, and consumer apps lean this way.

The chart shows the categories, from both dictionaries, where the two groups differ most.

The median posting in both groups asks for the same five years of experience. What differs is the language the company reaches for: keeping a large organization aligned, or getting product out the door with AI in it.

Is AI required?

The question I get asked most often is whether AI is required.

In this dataset, 56% of companies mention AI, ML, or LLMs somewhere in the qualifications, required or nice-to-have, but that number hides a lot. To separate a requirement from a mention, I looked only at the required qualifications and counted a company when at least one line there reads as experience: building or shipping AI products or LLM-based features, or hands-on work with the underlying models and systems. By that rule, 38% of companies (232) ask for AI experience as a requirement. Another 7% list it only under nice-to-have. In 2.3% of companies the required section mentions AI only as an interest or a passing familiarity, and in 8.1% it is a bare mention I could not classify either way.

On the duty side, 35% of companies want the person to build AI products or features, and 15% explicitly expect them to use AI tools in their own work. So AI experience is a stated requirement in about four in ten postings, mostly as knowledge of the technology the product is built on, not as a way of working. And the share varies a lot by industry:

More than half of the SaaS, AI, security, HR tech, and consumer-app companies require it. Fewer than one in ten of the financial services, retail, and gaming companies do. Several industries at the bottom have only 11-23 companies each, so I would not read too much into a single percentage there.

And which tools or technologies get named? For most companies, none. 69% of postings mention none of the tracked tool or technology names in the role sections. Among the ones that do, analytics tools come first (17.5% of all companies, mostly SQL), then PM tools (10%), then AI tools (9%, 56 companies). One caveat before the chart: a name in a posting does not say whether the candidate is meant to use that tool, integrate with it, or manage a product built on it, so read these as technology names, not a PM toolbox.

Claude turned out to be the third most-named tool overall, ahead of Confluence and Figma, which surprised me. Of the 38 companies that name it, 36 are outside the AI and developer-tools industries, including security, fintech, SaaS, and consumer apps.

By vendor, Anthropic's Claude and Claude Code appear in 38 companies and the OpenAI family (ChatGPT, GPT, OpenAI, Codex) in 19. Beyond that the AI toolbox is thin: Cursor 3%, Copilot 1.7%, Replit 1.3%, v0 1.2%, and not a single mention of Windsurf, Bolt, or NotebookLM. In this corpus the "AI-native PM" is mostly described in generic terms, and the few companies that do name tools name Claude or a coding assistant.

How often does the word "AI" appear at all?

The requirement dictionary above treats AI as a category, with all the judgment that involves. I also wanted the simplest possible check: does the word "AI" appear anywhere in the posting, in any form? So I searched every posting for "AI" in capitals, "A.I.", or "artificial intelligence," ignoring the lowercase pair inside other words and names like OpenAI.

• 32% of companies (195 of 605) never use the word "AI" anywhere in the posting.
• 42% do not use it in the role itself, meaning the duties or the requirements.
• One in ten postings (9.8%) mentions AI only outside the role: in the company blurb, the benefits, or a clause about using AI in recruitment.
• Counting ML, LLM, GenAI, and GPT as well moves the "never" share only from 32% to 30%.

The word count and the dictionary agree on 89% of postings (the dictionary finds AI in 61% of companies, the plain word in 68%), so the two measures tell the same story from different angles. By industry, the share of postings with no "AI" anywhere looks like this:

The industries where the word is absent are the same ones where AI was not a requirement: traditional financial services (83%), gaming (76%), retail (69%), and crypto (64%) mostly do not mention it at all, while no AI company and only 8.5% of SaaS companies skip it. Remote roles use the word more than located ones: 18% of remote postings never mention AI, against 35% of those located in Europe and 36% of those in the US and elsewhere. In industries with 11-17 companies one company moves the share by 6-9 points, so I trust the two ends of this chart more than the order in the middle.

How many years of experience?

Finally, the number I always check first in a posting: years of experience. In 78% of the postings there is a threshold; the other 22% say nothing about years at all. Among the 441 companies that give a product-specific number (another 31 give years without tying them to product work), the distribution is narrow.

Five years is the answer in 45% of them, and 4-5 years covers 55%. Only 4% ask for ten or more.

European postings skip the number far more often: 39% of the 158 postings located in Europe give none, against 15% of the 283 located in the US and elsewhere and 18% of the 164 fully remote roles. And postings that require AI ask for the same median of five years as those that do not.

Per industry, the median is five almost everywhere. The table shows where it is not, and who tends to skip the number.

Industrial companies ask for the most (median 7.5, and three of eleven want ten or more), traditional financial services sit at six, consulting at four. Crypto (50%), gaming (47%), consulting (39%), and financial services (39%) skip the number most often; proptech never does. Several rows have only 11-20 companies, so I read these as hints rather than rules.

What I take from it

What I take from all this, starting with what I am most sure of:

• The senior PM posting describes a coordination job and asks for proof you have done it before. The five most common duties (cross-functional work, roadmap, engineering and design collaboration, strategy, prioritization) appear together in 31% of postings, and the top requirements are years (78%), prior PM experience (73%), and a track record (65%). The advertised core still looks familiar. Whatever happens by 2030, this is what companies write down today.
• AI has entered the requirements as domain knowledge, not as a way of working. AI experience is a stated requirement in 38% of companies, a duty to build AI features in 35%, use of AI tools in daily work in 15%, and a named AI tool or technology in 9%. Whatever PMs do in practice (Productside reports that 4 in 5 of its 250+ respondents use AI daily), only 15% of postings put it in the job description.
• There are two contrasting bundles of language inside one shared core. Beyond the common five duties, half of the companies lean toward alignment, business cases, and Agile; the other half toward shipping, AI features, and product sense. Industry lines up with the side to some degree, but the split is too weak to call it a cause.
• Five years is the standard, with or without AI. 45% of the companies that state a threshold ask for five, and postings with an AI requirement have the same median.
• Most postings name no tool or technology at all. 69% mention none of the tracked names in the role sections; where one appears, it is SQL, Jira, or Claude.

One more thing to keep in mind about what a posting can tell us. When companies write a job ad, they often start from other companies' ads, so the same requirements, duties, and nice-to-haves travel from posting to posting whether or not the job behind them is the same. Some of the uniformity above may be that copying rather than the role itself. I see this piece as a contribution to the discussion about what a Senior Product Manager is expected to be, and as a record in time. If we come back to this and similar analyses in a few years, we should be able to see which way the role has moved, if at all: whether the emphasis in requirements and duties has shifted, whether new tools have entered the postings, and if so, which ones.

All of this is a snapshot of one dataset, mostly US and European tech companies, in one season. It can't show a trend, and it cannot show what the job is like once you are in it. It shows the job as advertised, nothing more.

If you have hired or been hired for a senior PM role this year, I am curious: did the posting look like the recurring core above, or like one of the two tendencies? And which of the requirements actually mattered in the interview?

Thanks for reading!

Methodology in brief

• Dataset: 605 Senior Product Manager postings, one per company, collected June-September 2026 from the author's job-postings database (488 sources).
• Classification: each posting split into duties, requirements, and nice-to-have sections by its headings, then matched against dictionaries of 33 duty categories, 41 requirement categories, and 55 named tools. The dictionaries were built and corrected by reading postings end to end, by the author and by two rounds of reviewers.
• Shares are the percentage of companies mentioning a category in the relevant section. Whole-text matching can add false positives and dictionary gaps can add false negatives, so treat them as directional estimates rather than exact values.
• AI experience signal: pattern match on the required-qualifications section only, with nice-to-have counted separately; for postings without detected sections, the whole text.
• Years: the lower bound of the stated range, product-specific lines preferred; industry labels from the author's own company-to-industry dictionary; industry cuts limited to industries with at least ten companies.

Sources

• Lenny's Newsletter, "State of the product job market in early 2026" (Lenny Rachitsky), March 2026.
• Products That Count and Mighty Capital, "2026 CPO Insights Report Predicts Product Managers Will Disappear by 2030", press release, May 2026.
• Productside, "The State of AI for Product Management Report 2026", 2026.
• Author's own analysis of 605 Senior Product Manager postings, one per company, snapshot September 2026.