No, AI Can’t Replace Product Managers Anytime Soon
And thinking it can demonstrates a profound misunderstanding of what we do.
The tech world has been shaken by the incredible number of layoffs in the last 18 months, and it seems like the majority of roles being slashed are product managers. Some thought leaders and pundits claim that this is the beginning of AI taking our jobs. Since I’m not in those conversations, I can’t say for sure whether leaders believe or hope that or not, but I do find it highly unlikely to be an effective strategy. Product managers cannot be replaced by AI for a few very explicit reasons, and thinking we can be demonstrates a profound misunderstanding of what generative AI is and can do (and what value those roles actually provide).
AI answers questions and performs simple tasks—but there still needs to be a person to ask the questions or assign the tasks.
AI cannot interpret its own responses.
AI cannot be creative in the ways we are.
“If I’d asked people what they wanted, they would have told me “a faster horse.””
AI can create versions of things we already have, but it cannot think outside the box because it IS the box. AI is literally the sum of all the work that has been done to date, with a powerful cross- and self-referencing system a human brain can only aspire to. But it takes a human—and usually a group of humans doing something mindless, annoying, or inspirational—to really have an idea so stupid or profound that it revolutionizes the way we live.
4. AI performs specific tasks—up to a point. It wavers when it runs into prolonged and complex trees of decisions and protocol.
5. AI cannot persuade humans using relational and social pressure.
Product Managers Jobs - Synthesis and Persuasion
(A product manager in the wild.)
A product manager’s primary job lies in the field of research and analysis. They answer the questions “What are we going to do?” and “Why do we want to do that?”
Understand business goals and priorities (communication and synthesis)
Ideate solutions that will move the needle towards the goals (research and analysis)
Validate the solutions’ viability (analysis)
Validate the solutions’ impact (analysis)
Prioritize the solutions based on viability and impact (analysis) → Roadmapping!
Socialize the proposed solution(s) (communication and influence)
Craft the actual solution (creative—design/writing)
Shape and iterate based on the feedback loops with internal stakeholders (research and analysis)
Remind all the squirrels of all the steps we just completed and are still working to execute. Daily.
Whether business leaders like it or not, what we do is actually quite complex. Also, like it or not, step 6 (communicating and exercising influence without official authority) takes up the majority of our time. No one would love that to go away more than product managers, believe me, but gaining alignment and keeping everyone on course in chaotic organizations with shifting priorities and distracted leaders and colleagues is not the easiest thing in the world. AI can’t do that, either.
In fact, step 6 is The Work. It infuses literally every step of the process in the majority of organizations because (1) we have to make sure our managers and leaders know that we’re busy working and moving things forward, and (2) we have to get approval every step of the way because we have no official authority and often are not decision makers.
Cue The Big Bang Theory:
Penny: What did you do today?
Leonard: I’m a theoretical physicist: I thought about things.
Penny: That’s it?
Leonard: I wrote some of it down.
And so many leaders don’t believe this is “real” work or that a computer program can’t replace it. Or that it needs to be done at all (that one always gets to me). If you’ve ever had to do this kind of work, thinking it’s not work is completely bonkers nuts. So a great big bear hug to every product manager who has ever been busier than a field of honey bees and still had to defend their progress and productivity to people who don’t understand their job function.
{{HUG}}
This is Not Like The Industrial Revolution
Source: USDA
We’ve all seen this chart that shows how jobs in agriculture performed a dramatic nosedive as the industrial revolution happened. I’ve seen a ton of people waving their arms in the air saying how AI is the same and we’re all going to lose our jobs and oh, the humanity.
Automation in agriculture is entirely different than automation and AI in knowledge work. Agriculture is a simple, physical process that can be automated to increase output in a quantifiable way. Knowledge work is subjective, the inputs and outputs are more complex and nuanced, and altering subtle inputs does not necessarily affect the outputs in a measurable way. (If a human can’t do it on command, don’t expect a machine to be able to since we are the ones operating the machines.)
This happened over the course of 200 years. Yes, it was a steady decline, and yes, I do believe that there will be fewer jobs in product management because our work can be supplemented by AI systems and tools.
On a team of 100 product managers, you could not replace 99 of them with AI tools and expect the same or a better outcome. You MIGHT be able to replace half—eventually. But certainly not today, and likely not in the next 10 years.
Welcome to the 21st century: the world of specialization.
Would it be easier, faster, and cheaper if this were not so? Sure. Is it plausible to get rid of this class of workers? Not in large, complex organizations. Like it or not, humans often require communicators and thinkers in various capacities because we are, well, human, and not everyone is good at communicating.
But by all means, let’s talk about how communicating is a “soft skill” and not really worth very much money. It’s a little thing, but there’s literally nothing without it.
So where do we go from here?
For the foreseeable future, this work needs to be done, there are people who specialize in it (hey, look, that’s what I happen to do), and AI cannot replace them. (Sorry about that.)
A message to leaders out there
Communication, influence, group organization, original research, contextualized analysis and synthesis is work that AI cannot replace. Don’t make the mistake of believing you don’t need these roles or that AI can compete with these humans, because it can’t. You will lose your edge if you try, and by the time you realize it, you’ll be playing catch-up. Yes, there is a balance to be found within your teams so they’re not dominated by these roles, and execution is streamlined, but you do need them.
A message to all our PMs out there
Do what you do best: document and communicate your case and your value. Learn how to use AI to increase YOUR value, upskill, and continue to crush it. We got this.
3 Ways to Work With Us:
Weekly Office Hours (free): Sign up for a private 30-minute meeting with our experts where you can ask any questions you have about product work or A/B testing. No obligation, no hassle. Just free collaboration. (Hey, we said this is why we get up, right? We meant it. Let’s talk.)
Quick Assessment (one-time fee): Sign up for a quick, one-time assessment of your website or app. Deliverables include a screen recording of the walkthrough, annotated screenshots, and our usability score and checklist.
Full Program Engagement: Our experts take you through the process of creating and starting a robust A/B testing program to achieve your goals, and all we ask of you is 1 hour a week.
