Lenny's Podcast: Product | Career | Growth

First principles: What it takes to be a great product manager in the AI era | Sherif Mansour

· Summarized

Share this episode
Anyone with this link can open it.
Share Link
Start at 0:00

Key takeaways

  • A PM’s central job is to fill customer, product-and-market, and business-context gaps so the team develops a shared brain and can make decisions without waiting for the PM.
  • In the AI era, throughput is becoming cheaper; direction, alignment, quality, and decision-making velocity are more important bottlenecks for established organizations.
  • The PM craft triangle—general manager, scientist, and artist—helps teams hire, calibrate, and assign PMs according to the work’s needs rather than treating all PMs as interchangeable.
  • Most PMs should build around their natural strengths and create complementary teams instead of trying to become exceptional at every dimension of product work.
  • AI should be applied first to a specific context gap, not adopted merely because PMs are expected to code or prototype; tool access does not replace domain expertise or design taste.
  • Juniors remain valuable because they bring AI-native experimentation and new ways of working, while experienced specialists provide domain judgment and detect AI-generated quality problems.
  • Starting with a solution is appropriate when observing customer workarounds, productizing repeated canvas behavior, or using existing company assets to shape strategy; it is not a universal replacement for problem discovery.
  • New managers should not become hands-off too quickly, and PMs working with executives should clarify context, decision ownership, and the strength of each stakeholder’s opinion before moving the team forward.

Summary

A great PM builds a shared brain

Sherif Mansour says the most important job of a PM in a medium or large organization is “filling context,” not personally making every decision. PMs should build a shared brain across their cross-functional team by continually supplying three kinds of context: customer, product-and-market, and business context. The goal is to close whatever gap is slowing the team—perhaps a vision gap this year, a customer gap next quarter, or a landscape gap later. He tells PMs to stay away from the Jira backlog and ask why they believe one card matters more than another, then give engineers, designers, and feature leads enough context to make those judgments themselves.

Context-sharing works best when the team experiences it

Mansour recalls complaining to a manager that too many decisions were blocked on him; she told him to spend more time on strategic work. He learned that scale comes from creating shared understanding rather than issuing instructions. He tries not to conduct customer interviews without engineers, because firsthand empathy and aha moments are stronger than receiving an AI-generated summary. Teams should also participate in visioning, market analysis, and experiments together. He gives examples of a PM having every engineer try different AI tools and Atlassian taking engineers to customer booths, so they hear feedback directly instead of receiving reports after someone else has “come down from the mountain.”

The Confluence checkbox came from an engineer’s customer insight

Mansour contrasts effective co-creation with simply handing over requirements. Designers should explore concepts with the team before polishing the final design, because participation changes everyone’s aha moments. His clearest example comes from 2010, when Atlassian noticed Confluence customers placing red and green star icons in meeting notes: red meant unfinished and green meant completed. Instead of adding more icons, an engineer asked why Confluence could not provide a checkbox inside a page. At the time, putting tasks into a document editor was not an established web pattern, but her direct observation of customers led to a pioneering feature. Mansour says this happens when engineers become “mini product managers” in their areas.

The PM craft triangle: general manager, scientist, and artist

Mansour describes Joff Redfern’s PM craft triangle, whose three points are general manager, scientist, and artist. A scientist PM naturally uses hypotheses, experiments, and A/B tests, making the type useful on growth teams. A general-manager PM thinks in total addressable market, revenue, profit, and business opportunities. An artist PM is emotionally and visually led, often beginning with an experience or vision and working backward toward the problem. The framework helps companies avoid throwing a generic PM into every team: hiring and calibration should ask what kind of work the team needs and how a candidate’s natural approach complements the existing group.

PMs should run toward their strengths

Mansour says most people cannot perform all three PM archetypes at a high level, though many can operate at a basic level and are valuable in zero-to-one work or startups. He openly jokes that he is Atlassian’s worst general-manager type and says this public admission creates safety for others to disclose weaknesses. Rather than spending a career trying to eliminate every weakness, he recommends discovering the work one is naturally good at and “running for the hills” toward it, then building a complementary team. He describes himself as a strong artist PM and says his colleague Jamil is nearly the opposite, making them effective together. A baseline of competence remains necessary, but nobody must be excellent at everything.

Career growth comes from the right opportunity at the right time

For early-career PMs, Mansour recommends looking for work where others have already praised the person’s contribution, then analyzing how they approached it and finding similar projects. A useful manager conversation is to show where one sits in the PM triangle and explicitly request more opportunities of that kind. For senior PMs, acceleration may mean pursuing a larger version of the same specialty or deliberately trying something different. Mansour recalls stepping aside so Rachel could lead a growing product because its scientist-oriented work suited her better. He also cites a colleague who has spent ten years at Atlassian becoming the company’s pricing-and-packaging specialist by continually finding work aligned with that strength.

Atlassian’s AI-builder experiment found a direction problem

Mansour challenges the recent narrative that everyone should become an AI builder and that PM and design roles are merging away. Atlassian moved PMs and designers onto experimental squads where everyone built with AI and there was no dedicated PM or designer. At first, one PM said the team was moving faster than ever. After roughly a month, he reported that everyone had returned to traditional roles because the experiment revealed that throughput was not the bottleneck—direction was. Mansour’s internal phrase was “direct the ship, don’t row it.” In a large organization with many engineers, faster output matters less than having people determine which direction the ship should take.

AI amplifies strengths, but specialization still matters

Mansour argues that PMs should own decision-making velocity rather than decision-making itself. Their first question about AI should be which context gap is slowing the team: prototyping may clarify an unclear vision, while agents may analyze customer or market information. Shipping production code is usually not the bottleneck unless the PM is in a startup, a zero-to-one project, or an unusually small team. He compares AI prototyping with the earlier Balsamiq mockup era: giving everyone a tool did not make everyone a good solution designer. PMs are often “some of the worst solution designers,” because tool access is not domain expertise. AI amplifies both strengths and weaknesses, so specialists remain essential for quality and taste.

AI-era hiring favors taste, experimentation, and complementary teams

Mansour says experience-led, artist-type PMs are increasingly valuable because when everyone has access to similar intelligence, taste and the quality of the experience become differentiators. He also looks for candidates who introduce new ways of working, especially people who can transform established teams’ habits around AI rather than merely saying they use it for summaries. He pushes back on the belief that established companies no longer need juniors. Younger hires may lack domain judgment and miss AI slop, but senior specialists are good at spotting quality problems while juniors bring experimentation and unfamiliar workflows. Atlassian’s AI-builder weeks showed juniors producing novel demos and seniors focusing on evaluation and quality; both contributions were needed.

Why PM roles may be thriving while other tech roles flatten

The host reports that his tech-job-market data showed PMs with the most open roles, while engineering openings were slowing, design was flat, and user research had only 108 open global roles at tech companies. Mansour connects the PM result to AI making throughput cheaper and creating more ideas, prototypes, and chaos. Organizations therefore need more people to establish direction, alignment, buy-in, quality standards, and goal fit. He offers a similar lesson from Balsamiq: democratizing mockups did not remove the need for design expertise. In an AI-native market, the circles of PM, engineering, and design capability may all grow without overlapping more; specialization and experience still determine the quality of what gets built.

Starting with a solution can be the right discovery tactic

Mansour does not reject the standard advice to begin with a customer problem, but says starting with a solution is useful in specific contexts. A general canvas can reveal what customers want to productize: Atlassian observed customers using RovoChat to group Confluence whiteboard sticky notes, then turned that awkward chat interaction into a native feature for selecting notes and grouping them into themes. Similarly, customers place Rovo agents into Jira and Confluence workflows; Atlassian can leave some uses infinitely customizable while turning repeated patterns into native capabilities such as AI triage. He calls this “paving the cow path”: observe a manual customer solution, then make that path easier.

Solutions also anchor product strategy

The second legitimate reason to start with a solution is product strategy. Mansour says strategy must answer where a company plays and how it wins, which requires examining the assets and capabilities it already has. He uses Amazon’s bookstore as an example: Jeff Bezos began with the solution of an extraordinarily large inventory, then recognized that the same asset could support an everything store. Mansour advises teams to list their existing assets and their beliefs about the future, then look for connections. He still considers “what problem are we solving?” a valuable question, but argues that teams should understand when an existing customer workaround, product capability, or strategic asset makes solution-first exploration more productive.

Established products should often bolt AI onto existing workflows

Mansour calls “don’t bolt AI on; reimagine the workflow” a useful but incomplete slogan. For an established product with millions of users, he recommends studying existing human workflows and adding AI where it can generate value, improve efficiency, or assist a step. Atlassian found that customers had already modeled billions of workflows in Jira, including non-software processes, so it added agents to columns, cards, and workflow loops rather than rewriting the product from scratch. This lets customers move from AI novice to AI native gradually. Mansour says an AI-native product might have been designed the same way: define a workflow and let agents work on portions of it. “Bolting on” can therefore be the practical path to transformation.

Managers, individual contributors, and visionary stakeholders

Mansour rejects the idea that every senior PM must become a manager to participate in strategy. Atlassian supports dual career tracks, with distinguished individual contributors contributing at executive level; people can also switch tracks depending on the season, such as managing during a growth phase and returning to IC work for a zero-to-one effort. For new managers, his main warning is not to become hands-off too quickly. Strong managers create space but also enter the work to help teams reach outcomes. When working with visionary founders, he imagines the leader being on chapter ten of a movie while the team is on chapter one. He advises filling the context gap before simply following orders.

Disagreement requires explicit ownership and balanced arguments

When Mansour catches up to a founder or executive’s reasoning and still disagrees, he states his case and asks how strongly the leader feels and who owns the decision. Leaders often respond that they disagree but want the PM to decide, and Mansour will support that decision. When communicating a choice to the team, he presents the pros and cons of both options, explains why one is superior, and acknowledges legitimate trade-offs without unnecessarily naming the individual who preferred the alternative. He urges PMs to argue against their own proposal and ask, “Why would we not do this?” Before acting on a powerful stakeholder’s opinion, the PM should clarify whether it is information, a suggestion, a recommendation, or a firm decision.

Sponsors

  • WorkOS · WorkOS provides drop-in APIs for enterprise features including SSO, SCIM, RBAC, and audit logs. The host also described its agent capabilities: Agent Auth for scoped, revocable identities; agent registration with AuthKit; and Airlock for allowing, denying, or escalating agent actions to humans. He said many startups expanding upmarket use WorkOS and encouraged listeners to make their apps enterprise-ready.
  • DX · DX measures and benchmarks the impact of AI adoption across the product lifecycle. It shows where AI helps developers or creates bottlenecks, evaluates AI vendors, monitors costs and licenses, and identifies what limits engineering agents. The host said companies including Snowflake, Sony, and BNY use the platform and directed listeners to request a demo.

Show notes

Sherif Mansour is Head of AI and Product Craft at Atlassian (the company behind Jira and Confluence), where he has spent 17 years building and scaling some of the most successful enterprise products in history. He joined Atlassian as its first PM on Confluence, eventually becoming its first Distinguished PM, and now leads both the craft of product management across the company and Atlassian’s AI product strategy.

We discuss:

1. Why a PM’s #1 job is to build a “shared brain”

2. The artist, the scientist, and the GM: three types of PM, and how to tell which one you are

3. Why PM roles are growing fastest in the AI era

4 Why Atlassian is hiring more juniors than ever

5. The most common mistake new PM managers make

—

Brought to you by:

WorkOS—Make your app enterprise-ready, with SSO, SCIM, RBAC, and more

DX—Engineering intelligence for the AI era

—

Episode transcript: https://www.lennysnewsletter.com/p/first-principles-what-it-takes-to

—

Archive of all Lenny's Podcast transcripts: https://www.dropbox.com/scl/fo/yxi4s2w998p1gvtpu4193/AMdNPR8AOw0lMklwtnC0TrQ?rlkey=j06x0nipoti519e0xgm23zsn9&st=ahz0fj11&dl=0

—

Highlights: https://lennyspodcast.com/mostreplayedmoments

—

Where to find Sherif Mansour:

• X: https://x.com/sherifmansour

• LinkedIn: https://www.linkedin.com/in/sherifmansour

• Medium: https://sherifmansour.medium.com

• YouTube: https://www.youtube.com/@sherifmansour8057

—

Where to find Lenny:

• Newsletter: https://www.lennysnewsletter.com

• X: https://twitter.com/lennysan

• LinkedIn: https://www.linkedin.com/in/lennyrachitsky/

—

In this episode, we cover:

(00:00) Teaser

(01:24) What makes a great PM

(04:17) Why your job is to fill context gaps, not make decisions

(07:10) Examples of context-sharing done well and done poorly

(11:25) The GM/scientist/artist PM triangle

(12:38) How to use the triangle in hiring and calibration

(15:01) You don’t need to be great at everything

(18:09) Career coaching: early-career unlocks

(21:14) Career coaching: senior PM and manager unlocks

(22:19) Why PMs are thriving in the AI era

(23:42) Atlassian’s AI builder experiment—and what it revealed

(25:01) Why throughput is no longer the bottleneck; direction is

(28:48) What PMs should actually be doing with AI

(31:04) What’s trending up and down in PM hiring

(32:44) The case for hiring more juniors in an AI world

(38:07) How to spot the artist streak early in candidates

(46:05) When starting with the solution is the right move

(52:42) Why “bolting AI on” is often the right product strategy

(56:01) The importance of nuance over hot takes

(01:01:31) Why managers still matter

(01:02:42) The most common mistake new PM managers make

(01:04:05) How to work with visionary founders and strong stakeholders

(01:07:12) When you disagree: a framework for transparency

(01:10:51) Final thoughts

(01:11:37) Lightning round

—

Referenced:

• Atlassian: https://www.atlassian.com

• Jira: https://www.atlassian.com/software/jira

• Confluence: https://www.atlassian.com/software/confluence

• Product managers make all the decisions, right?: https://sherifmansour.medium.com/product-managers-make-all-the-decisions-right-87ea80beb37a

• Joff Redfern on LinkedIn: https://www.linkedin.com/in/mejoff

• How to: Use the Product Management Craft Triangle: https://sherifmansour.medium.com/how-to-use-the-product-management-craft-triangle-a6048b1c9bc9

• Jamil Valliani on LinkedIn: https://www.linkedin.com/in/jamil-valliani-b131881

• Mike Cannon-Brookes on LinkedIn: https://www.linkedin.com/in/mcannonbrookes

• State of the tech job market: late 2026: https://www.lennysnewsletter.com/p/state-of-the-tech-job-market-late

• Direct the ship, don't row it: https://sherifmansour.medium.com/direct-the-ship-dont-row-it-28785929626f

• Balsamiq: https://balsamiq.com

• On hiring "juniors" in the world of AI: https://sherifmansour.medium.com/on-hiring-juniors-in-the-world-of-ai-b1d05843f167

• Product engineers: https://sherifmansour.medium.com/product-engineers-f424da766871

• Lenny’s MCP: https://www.lennysdata.com

• Noam Segal on LinkedIn: https://www.linkedin.com/in/noamsegal

• Why Atlassian has dual track product management career paths and how we did it: https://sherifmansour.medium.com/why-atlassian-has-dual-track-product-management-career-paths-and-how-we-did-it-7276ecaa6eea

• FlashTags: A Simple Hack for Conveying Context Without Confusion: https://www.onstartups.com/flashtags-a-simple-hack-for-conveying-context-without-confusion

• Zigging vs. zagging: How HubSpot built a $30B company | Dharmesh Shah (co-founder/CTO): https://www.lennysnewsletter.com/p/lessons-from-30-years-of-building

• Project Hail Mary on Apple TV: https://tv.apple.com/us/movie/project-hail-mary/umc.cmc.7jxdlxvz304lj3iwhtrhbe8fv

• Google Maps navigation gets a powerful boost with Gemini: https://blog.google/products-and-platforms/products/maps/gemini-navigation-features-landmark-lens

• 100+ Atlassian T-shirts from the last 12 years: https://www.youtube.com/watch?v=zsE15RWZwFY

—

Recommended books:

• Invent and Wander: The Collected Writings of Jeff Bezos: https://www.amazon.com/dp/1647820715

• The Prison Healer: https://www.amazon.com/dp/035866943X

• On the Edge of the Dark Sea of Darkness (The Wingfeather Saga, Book 1): https://www.amazon.com/clp/0525653562

—

Production and marketing by https://penname.co/. For inquiries about sponsoring the podcast, email [email protected].

—

Lenny may be an investor in the companies discussed.



To hear more, visit www.lennysnewsletter.com

Follow this show on AudioBrevity and new episodes get summarized for you. Create a free account

Couldn't load the audio. A blocker may be stopping the show's tracking links. Open the file