Product Managers Face AI Identity Crisis
Key Points
- AI is creating a uniquely severe identity crisis for product managers, more disruptive than the impacts felt in engineering, sales, marketing, or customer success.
- While other functions have predictable AI roles—CS for ticket triage, sales for call coaching and email drafting, marketing for creative assets—PMs confront multiple, overlapping threats across product definition, insight generation, and execution.
- AI can automate many PM deliverables such as PRDs and other documentation, but delivering quality still demands deep product sense, storytelling, and precise prompting to retain thoughtful, creative output.
- Building AI‑driven products adds a layer of complexity because these products are inherently probabilistic, requiring PMs to manage uncertain behavior and edge‑case scenarios unlike traditional software.
- Executives often push for rapid AI releases, while engineering teams are hesitant to commit to fixed deadlines on probabilistic systems, creating tension and further compounding the PM crisis.
Sections
- AI Sparks Identity Crisis for PMs - AI is rapidly assuming core product‑management functions—like PRD drafting and strategic decision support—creating a multifaceted threat that leaves product managers uniquely unsettled compared to the more predictable AI impacts seen in sales, marketing, and customer success.
- PM Role: Glue Amid AI - The speaker argues that product managers have historically served as the connective glue between engineers and stakeholders, yet AI advancements and expanding duties—such as prototyping and even coding—are blurring and stressing this already ambiguous role.
- Finding Meaning in Product Management - The speaker urges PMs to use AI tools like ChatGPT for learning, prioritize projects that genuinely move the needle, and secure autonomy to prevent burnout and maintain motivation.
- Human Conviction Over AI Automation - The speaker stresses that AI can assist with drafting documents and prototyping, but it cannot replace the leader’s personal judgment, stakeholder alignment, and responsibility in guiding projects and meetings.
Full Transcript
# Product Managers Face AI Identity Crisis **Source:** [https://www.youtube.com/watch?v=nQcy-YlYpng](https://www.youtube.com/watch?v=nQcy-YlYpng) **Duration:** 00:12:44 ## Summary - AI is creating a uniquely severe identity crisis for product managers, more disruptive than the impacts felt in engineering, sales, marketing, or customer success. - While other functions have predictable AI roles—CS for ticket triage, sales for call coaching and email drafting, marketing for creative assets—PMs confront multiple, overlapping threats across product definition, insight generation, and execution. - AI can automate many PM deliverables such as PRDs and other documentation, but delivering quality still demands deep product sense, storytelling, and precise prompting to retain thoughtful, creative output. - Building AI‑driven products adds a layer of complexity because these products are inherently probabilistic, requiring PMs to manage uncertain behavior and edge‑case scenarios unlike traditional software. - Executives often push for rapid AI releases, while engineering teams are hesitant to commit to fixed deadlines on probabilistic systems, creating tension and further compounding the PM crisis. ## Sections - [00:00:00](https://www.youtube.com/watch?v=nQcy-YlYpng&t=0s) **AI Sparks Identity Crisis for PMs** - AI is rapidly assuming core product‑management functions—like PRD drafting and strategic decision support—creating a multifaceted threat that leaves product managers uniquely unsettled compared to the more predictable AI impacts seen in sales, marketing, and customer success. - [00:03:10](https://www.youtube.com/watch?v=nQcy-YlYpng&t=190s) **PM Role: Glue Amid AI** - The speaker argues that product managers have historically served as the connective glue between engineers and stakeholders, yet AI advancements and expanding duties—such as prototyping and even coding—are blurring and stressing this already ambiguous role. - [00:06:29](https://www.youtube.com/watch?v=nQcy-YlYpng&t=389s) **Finding Meaning in Product Management** - The speaker urges PMs to use AI tools like ChatGPT for learning, prioritize projects that genuinely move the needle, and secure autonomy to prevent burnout and maintain motivation. - [00:10:14](https://www.youtube.com/watch?v=nQcy-YlYpng&t=614s) **Human Conviction Over AI Automation** - The speaker stresses that AI can assist with drafting documents and prototyping, but it cannot replace the leader’s personal judgment, stakeholder alignment, and responsibility in guiding projects and meetings. ## Full Transcript
AI is causing a crisis for product
managers. I don't think it's too far to
say that PMs are the worst off of the
job families around AI right now. And I
say that being very aware that there are
colleagues of mine in engineering, in
customer success, in sales who would
disagree with me in marketing who would
disagree with me. That's okay. I want to
talk about PM partly because I know it
well. Done PM. I've managed PMs. I've
led PMs. But partly because AI is doing
more of the heavy lifting in the PM
domain than most PMs expected and that
is leading to a crisis of identity that
is distinct and different from what I
see when I talk to marketers to CS to
sales to others who feel like their jobs
are impacted by AI. Let me explain what
I mean. If you're in CS, it's a very
clear engagement model for AI. The AI
comes and whatever your feelings about
it, it's picking up triage tickets.
That's just pretty much what it does. If
you're in sales, the AI comes and it's
listening in on calls and it's giving
you coaching tips and it's maybe writing
your out outbound emails, etc. Again,
very predictable scope of engagement. If
you are in marketing, the scope of
engagement is similarly predictable, but
it's about creative and asset
generation. It is different with PM. And
you guys give the PM you know a hug
because for PMs there's multiple threats
on multiple axes. Yes, we have the same
thing that you guys have where it's
about asset generation. So PRD
generation, Clarvo has got a whole
business around PRD generation and
there's other assets that PMs are
responsible for that AI helps with as
well. So there's that whole skill set to
master and by the way that is not an
easy skill set because you have to
retain your thinking, thoughtfulness and
creativity but somehow work faster
because of AI. There's product insight
requires a mixture of data and
storytelling genuine product sense
product gut engineering competence. It
was always hard and AI makes it faster
but not necessarily better unless you
really know how to prompt. So that that
whole asset creation thing is very
fraught for PMS but on top of that you
have to build AI product which marketers
don't have to do CS doesn't have to do.
You have to figure out how to build AI
product and AI product doesn't work like
any other product guys. other products
you can just sort of say this is what
the product is and write the
requirements and that's how we were all
brought up if we were npm for a decade
but not anymore now the product is
probabilistic if you're building an AI
product the product is it mostly does
this but sometimes there are edge cases
sometimes the product is we figured this
out can you help us package it is not
easy to build AI product especially when
you couple that with the demands from
executives where executives are
expecting really rapid ships on AI.
They're expecting deadlines and your
engineering team is like not willing to
give you deadlines on probabilistic
products so often. So there's the
product piece, there's the asset piece.
Both of those are very tricky. There's a
third piece that no other role has,
which is that PM has always been a glue
role. It has always been an in between
role. In fact, it evolved out of the
need to keep engineers out of meetings,
which is why if your PM is in meetings,
he or she is they're doing their job.
But so much of that is now up for
debate, isn't it? Because if it's just
about giving information, can an AI do
that? Yeah. If it's just about
stakeholder management, can an AI do
that? Maybe. Not really. If it's about
persuading and aligning on a step
forward with executives, can an AI do
that? Definitely not. And so that
there's this weird stakeholder
management mess that PMs have always had
to handle that they now have to handle
with AI in ways that are not clear. And
on the other side, it's getting
increasingly blurry, too, because now
PMs are expected not just to write
specs, but to directly prototype. We're
expected to write stuff up in lovable
and show it. And maybe we don't write
the PRDM, we just show the the
prototype, right? Or maybe we go so far
as to commit code using claude code and
maybe an engineer reviews it or maybe
codeex reviews or maybe claude code
reviews it. But either way, we're
writing code. We're vibe coding code.
We're vibe coding our SQL statements.
Maybe we're making small UX pieces we
can just commit ourselves. PM has always
been an ambiguous role. The fact that
it's ambiguous is not new. The fact that
it is under strain from so many
different axes at once. The job itself
is changing both with stakeholders and
with engineering at the same time as the
tools we use to do the job are changing
at the same time as the definition of
the role itself is changing as the same
time as the products we build are
changing. That is why PM is in crisis.
That is why I know a lot of PMs who are
walking away because it is just too
hard. Don't worry, the first half of
this video is all about how bad it is.
But the second half is hope. The second
half is words of wisdom from someone who
knows AI pretty well and who knows PM
pretty well on how we start to navigate
through this as product managers, as
product leaders. Number one, we need to
start to get real comfortable with the
technical aspects of LLMs. Increasingly,
the career risk is hitting non-technical
PMs harder. And I will go farther. It is
hitting technical PMs who are not AI
technical fluent. And that is different
from saying I need to be a PM who can
manage an AI product release. I'm
talking about be a PM who understands
when to enforce schema validation or
not. When to have a library of tools
versus having a large prompt in your
agent. You want to be really familiar
with the tradeoffs that go with building
AI products. And in particular, you want
to have a highfidelity mental model of a
large language model and how it works,
how agents work so that you can be
helpful when you are in the room with
engineers talking about tradeoffs
because that is how you actually move
the ball forward on product and
eventually come back with relevant
deadlines. So I think the first piece is
you got to get technically fluent.
There's I'm sorry there's no substitute.
The good news is chat GPT is a great
teacher. Uh it really is like you can
ask ask chat GPT to give you a technical
AI lesson every morning and schedule it
and it will do that. Number two, make
sure that what you are building matters.
I say that because I think that another
reason why burnout happens and
frustration happens is that PMs are put
especially now on projects that are
unlikely to move the needle. And if you
think, oh that's the nonI products, I've
got news for you. AI products don't
always move the needle either. In many
cases, what burns PMs out is being asked
to do an AI product that the CEO came up
with on LinkedIn and they have no
autonomy to manage that and they just
have to push it through and good luck.
So, get yourself into a space where you
have a product that actually matters. It
might not be an AI product and I'm going
to tell you that that's okay. But it
needs to be a product that you believe
can move the needle. Even if it doesn't
yet, you believe your involvement is
enough to move the needle. And if that
sounds like it's meaning making, it is
meaning making. But that's the art of
PM. If you're not motivated, if you're
not energized, if you don't feel like
there's meaning here, you can't convince
your stakeholders. You can't sell the
product, you can't believe in the
product, you can't roadmap the product,
and you cannot convince engineers on
that. So yeah, it does. And I think it
especially matters in the age of AI
because so many of these AI products
that PMs are being asked to PM. They
tell me and I see are just AI washing.
They're they're not any good. And that's
so demoralizing. It doesn't help with
your resume. It doesn't build value. It
doesn't move the needle for customers.
It discourages your engineering team. It
doesn't teach you good AI best
practices. So build what matters. Number
three, make sure that you don't lose
your product intuition. So AI is going
to help you go faster. You're going to
be asked to go faster and you're going
to feel split between five, six
different ways. How many ways can I
write a Slack update? Can I use the AI
to write the PRD for me, etc. The best
PMs are basically treating AI as a way
to extend their attention on less
important tasks and they are keeping the
things that matter, the taste, the
judgment for themselves. They are not
outsourcing that. If you have a hunch
about your product and you're an
experienced PM, follow that hunch. Take
it seriously. Yes, AI can listen to and
read through all of the transcripts from
all of the customer calls, and that's
great. And maybe it will surface some
things you didn't see, and that's
amazing. You have a product gut for a
reason. And it is demoralizing to ignore
it. It's death to your product gut. It's
damaging to your future career. This is
this is a craft skill, right? It's a
fingertippy skill. That's something you
learn in the shop. Don't lose your
touch. Don't lose your product
intuition. If you think it needs to
launch now, it probably does. Just go
with it. And I think that we sometimes
misunderstand what AI is. We sometimes
think because AI is something we're
building something we're building with
and also like apparently an AI colleague
and also apparently involved in our
meetings and stakeholder management that
we should give it a lot more deference
or trust then I think we should
especially in product where you are
responsible for setting direction and
autonomy don't do that trust your gut
trust your intuition you will get less
burned out the product is going to be
better and you will put AI in its right
place as an assistant not a colleague
that calls the shots. I'll tell you one
more thing. Your job to drive alignment
between engineering and leadership has
not gone away, not eroded one bit. AI
can't help you with it. And if you can't
keep direction there and clarity there,
your career isn't worth a plug nickel.
That is the value. That is what you
can't substitute any other way. And so
make sure if you're working on something
meaningful that you are putting the
human work in to keep alignment with
leadership and your technical teams that
you can't have AI just write the deck
and assume it's going to be right there.
Right? You you can't have AI give the
presentation. You are going to have to
go in you are going to have to write the
docs and the decks that work for you.
And can AI help? Yeah, actually AI can
help. But there's no substitute for your
conviction that this is the right angle
to approach an important meeting with.
There's no substitute for your
conviction that the agenda needs to be
shorter because of X or Y stakeholder.
There's no substitute for your
conviction. The technical team needs two
weeks more time and you have to go and
get it. That's your job. And so when you
look back and you look at sort of where
you are now, the things that I'm
emphasizing are things that you can
trace a through line for the last decade
plus. These are things we've always had
to do, but I think they're things we've
started to forget because AI is
incredibly loud and in our faces all the
time. I'm not saying don't use AI. AI
tooling is great. If you can use AI to
help you write better PRDs and one
pagers and write better SQL and get into
prototyping, that is all great skills
that you can demonstrate that move the
ball forward for your customers. Do it.
Don't lean out, lean in. But you are
leaning into AI to become a PM with
these core skills that don't change like
working with stakeholders, like aligning
with leadership, like working on
something that matters in a way that's
new. AI is a tool in the toolkit. You
are a craftsman and you can use AI as a
tool in your toolkit and you should and
you need to and the ones who don't are
in trouble. But don't mistake that need
to lean in for a need to lean away from
these core values. Those don't change.
And I think that I feel like this is
missing. I feel like we don't have
enough conversations about the core
values in the job family. And that's why
I'm putting this out there. They haven't
changed. The things that make PM
successful over time have not changed.
AI is just a tool that we're using to
get there. I hope this has helped. Best
of luck with product. Well, if you
stayed this long, you want to check out
my product manager dice. I have I can
roll this, right? And it says on the
road map. It says no. It says plan for
Q5. That's my favorite one. And uh it
depends. I love that. So anyway, have
fun, keep chuckling and uh stay in
product is worth