For those following along… I lied. This blog is 100% generated by Claude (Fable 5.1, High). I got the PMM Jam session transcripts kind of late, so I used AI. The prompt I used is at the bottom of this blog entry. I didn’t edit a word of the output. Even though it created 1600 words when I asked for 1200.
An interesting question for you: Compared with the piece I wrote 100% myself yesterday, warts and all (and there are many), in ~3.0 hours which is better? Which one has more authenticity? Which one reads as more human, and “more Gary,” if you’ve read my stuff before?
It saved me time and will let me eat dinner with my wife. But would you rather read something written by me?
The AI interview question is not about AI. It never was.
Here is my challenger thesis for Day 2 of the Fall 2026 PMM Jam: the scariest new question in a product marketing interview (“So, how do you use AI?”) is the oldest question in the book wearing a new hat. The hat is new. The head underneath it is the same one hiring managers have been trying to see inside of since I interviewed for my first marketing job in the age of fax-back forms.
Jonathan Pipek, co-founder of 3AM Recruiting, kicked off the Jam with “What the AI Are Interviewers Looking For?” and he said it better than I just did: whenever a recruiter or hiring manager asks how you use AI, they’re really trying to understand how you work. Not which model. Not which subscription. How you work.
An antecedent from my own job-search archive. In 2011 I was unemployed and met Eric Yuan at an ed-tech show in Boston, back when Zoom was a few people and a big idea. I didn’t get an offer, and I’ve written elsewhere that the fault was mine: I showed up with my tools and my history, and I should have better understood what he was looking for at that point. Fifteen years later, the tools have changed and the mistake hasn’t. Jonathan’s whole hour was, in effect, a field guide to not making it again.
What Jonathan actually told us (the recorded hour, anyway)
Jonathan and his co-founder Yi Lin Pei are PMM leaders who recruit, not recruiters who dabble in PMM, and the session showed it. He opened by polling the room, and the answers were a tidy little horror anthology: “How do you use AI day to day?” “Have you launched an AI product?” “How do you review its output?” One attendee gave the honest answer (“I use it to draft content and produce more assets”), watched the interviewer’s face go flat, and didn’t get the job.
His fix is a structure, not a script. Give a 30-second overview of how AI shows up across your workflows, then go deep on one example, and lead with the outcome (“I built an ROI calculator in Lovable for sales; it contributed to a 15% lift in conversion”). Then walk six things: the problem, why AI instead of the old way, your approach, where your human judgment stepped in, the proof it worked, and the result.
Two things I appreciated most. First, the bar moves with your level. An IC should show how AI made them more capable. A director should show what changed for the team (what got adopted, who kept it current, what broke on day three and how it was rebuilt). Second, be tool-agnostic. He told a story about a strong candidate who spent an entire consulting interview on what they’d built in Claude for a client that could only run Copilot, and never explained the underlying thinking that would have transferred in an afternoon. His gardener line stuck with me: nobody cares whether you used a shovel or a digger. They care that you saw the rocky soil and knew which one to pick up.
(The second, unrecorded hour was the candid part. Per my own house rule, I’m not sharing it. Join the Jam and get it firsthand.)
We’ve been here before. Sort of.
Think about the decade before the AI boom, roughly 2013 to 2023. The resume keyword was the martech stack. Marketo certified. HubSpot certified. “Proficient in Salesforce.” Applicant tracking systems were screening for those nouns, so we all learned to stuff them in, and interviewers learned to ask “which tools have you used?” as a proxy for “can you do the job?” It was a lousy proxy then. Jonathan’s point is that “I use Claude” is the same lousy proxy with a 2026 timestamp, and the better interviewers have already stopped accepting it.
What’s the same: the underlying ask has always been judgment plus proof. Show me a problem, the call you made, and evidence it worked. That was true when the artifact was a printed brochure in a leather portfolio, and it’s true when the artifact is a Lovable prototype.
What’s different, and this is the part that matters: for the first time, the tool itself can manufacture a plausible-looking answer. Marketo never wrote your campaign brief for you. A first-draft battle card that “claims features that don’t exist” (Jonathan’s words, and he’s seen it) is new territory. So the interviewer’s question has quietly shifted from “can you operate the tool?” to “can I trust what you let out the door after the tool touched it?” That’s why his fourth element, human judgment, is the one I’d put in bold. It’s the only one the tool can’t fake on your behalf.
The other difference: the tools churn monthly, not yearly. Which makes the tool-agnostic framing less a nicety and more a survival skill. Nobody wants to re-interview every time a model gets renamed.
Gen Z and Gen X are in the same leaky boat
Here’s an uncomfortable symmetry. A 24-year-old with a marketing degree and a 54-year-old with twenty years of launches have the exact same problem walking into Jonathan’s six-element framework: neither one has a field-tested, in-industry AI example with proof attached.
The Gen Z candidate has the AI fluency and no company to have used it in. Their examples are class projects and side hustles. The Gen X (and, ahem, older) candidate has the company history, but it’s often history at a shop where AI was banned, or was Copilot-only, or arrived six months before the layoff. Their examples are weekend experiments. Sometimes very sophisticated weekend experiments. Still, no sales rep ever used them, so there’s no “proof” and no “result” to lead with.
Both groups are tempted by the same wrong answer: certifications. Jonathan was blunt that, past your first couple of years, a cert shows you took a course, not that you can do the work. Hiring managers want to see what you actually did.
And both get the same right answer from him, which I found generous: if the real thing is confidential, or never existed, recreate a small section with the sensitive details removed. Build the battle card for a made-up competitor. Annotate it: here’s what the AI drafted, here’s the claim I cut because the research didn’t support it, here’s how it got used (or how it would be used, say so plainly). The point of a portfolio piece isn’t that it shipped at a Fortune 500. The point is that a stranger can see your judgment on the page without you in the room to narrate it.
Which is exactly what a PMM Jam mission is for
Read that last paragraph again and then look at how Eve Horne designed this Jam. The theme is AI-agnostic: the underlying PMM skills that hold their value no matter which tools come and go. (I wrote about why that theme is the right one in my Electric Boogaloo post, and about the improv-jam origin of the name on kickoff day.) Jonathan’s first-principles advice and Eve’s theme are the same idea viewed from the recruiter’s chair and the practitioner’s chair.
Then there’s the mission. Every Jammer picks a real PMM challenge (positioning, launch, messaging, or a wild card), does the work over two weeks with mentor office hours, and submits it for judging. Finished missions can be published as a case study on Dusted.io, where other PMMs upvote them for the PMM Choice Award.
Map that onto Jonathan’s six elements and the mission fills in the two boxes that Gen Z and Gen X were both missing. Proof: a panel of senior PMM leaders reviewed it and gave feedback. Result: it was published, upvoted, maybe awarded, and it’s sitting on a public URL that a hiring manager can click before you’re even in the room. That’s not a certification. That’s work, with witnesses.
It also solves the “what do I show?” problem for the person whose real work is under NDA. The mission is yours. No employer IP, no anonymized screenshots, no awkward “I can’t really show you that one.” Annotate it the way Jonathan showed us, write the one-paragraph executive summary on top, and you have a portfolio piece that works in an interview, on LinkedIn, and on Dusted.
So my homework for myself this week, as a Gen X’er who has used these tools on more side projects than I’d care to admit and far fewer in-office ones: stop treating the mission as a class assignment. Treat it as the specific example I’ll lead with in my next interview. Outcome first.
Thank you, Jonathan
The slides are being shared with Jammers, and Jonathan and Yi Lin recently launched bestpmmjobs.com, a job board stocked with PMM roles, including ones 3AM is actively filling. If you’re looking, look there. If you’re hiring a first or second PMM at a seed-to-Series-B company, 3AM is built for exactly that.
It’s not too late to join the Jam. Recordings of the first hour of each session are available to participants, and on-demand access runs through January 1, 2027. The unrecorded second hour, though, you have to show up for. I’ll be in the Slack, and I’ll be here tomorrow with Day 3.
Am I full of it? Is “same question, new hat” too glib? Tell me in the comments. You can find me at gdietz@garydietz.com and on LinkedIn.
Sources: Jonathan Pipek, “What the AI Are Interviewers Looking For?”, PMM Jam Fall 2026, Nov. 9, 2026 (recorded hour and slides) · pmmjam.com · 3amrecruiting.com · Jonathan Pipek on LinkedIn · Dusted.io

Leave a Reply