At a Glance
This blog redefines product management for 2025. As AI accelerates how we research, design, and build, the role must shift back to commercial discipline, intentional decision-making, and human judgement. Product managers must slow down just enough to make smarter, more meaningful choices in a world obsessed with speed.
This piece has taken me over 20 years to write. Not because I haven’t tried, but because there are already countless articles attempting to define product management. And yet, it feels more urgent than ever to revisit the question because the profession is shifting under our feet. Product management is evolving again, and this time, the change is accelerated by something new: artificial intelligence.
If we aren’t clear about the role in 2025, we risk losing it entirely, being overtaken by automation, and misunderstood by executives who focus on our outputs instead of impact.
From Commerce To Delivery — And Back Again
When I started in product management in the early 2000s, the function was a combination of commercial and delivery. Our job was to find and grow profit. We were funded by the business and measured on commercial outcomes.
I remember assessing opportunities to sell dial-up internet in regional Australia. My analysis was backed by investment dollars. We were constantly asking: Who is our customer? Why will our customer buy our product? How do we differentiate? How can we scale it profitably?
Then (the commercial variation of) Agile arrived, and with it, a powerful focus on speed, iteration, and collaboration. It was transformative, but it also shifted product management away from its commercial roots and deep into delivery. We learned to ship fast, but somewhere along the way, we forgot to ask why.
As Agile spread, other sectors wanted in — government agencies, not-for-profits, corporates looking for “speed and innovation.” They built product teams and adopted agile rituals, but they rarely paused to ask what the role of product management really was. It was never about vision or strategy formation. It became about managing backlogs, not shaping direction.
We built momentum, but not always meaning.
Now, the pendulum is swinging back. The global economy has made capital more expensive, so product management is returning to its roots: commercial discipline, hard choices, and value creation. We’re expected to be stewards of investment again. As we should be.
But this time, we have a new force shaping how we think and work: AI.
A Brief History, A Long Debate
Product management isn’t new. It’s almost 100 years old, with roots in 1930s brand management and a long lineage through companies like Hewlett-Packard, Microsoft, and Google. Each generation has redefined what the role means: part strategist, part maker, part diplomat, (sometimes) part janitor.
We’ve also spent decades debating definitions. PDMA, ISPMA, and AIPMM tried to codify the field through the ProdBOK. It didn’t stick. Why? Maybe that’s because product management is so context-dependent. Or maybe it’s because, as a profession, we’re quietly stubborn, convinced that our way is the right way.
Still, the debate matters. Because clarity about what product management is also clarifies what it isn’t.
Four Ways To See Product Management
We talk about “product management” as though it’s one thing. It’s not. It’s four overlapping lenses:
- A Function — A formal team with roles, budgets, and accountability.
- A Set of Activities — Prioritising problems, experimenting, and aligning people around outcomes. Product thinking doesn’t need a title.
- A Body of Knowledge — Frameworks, theories, and tools for decision-making, prioritisation, and strategy.
- An Operating Model — A way an organisation works: customer-first, outcome-oriented, and adaptive.
When we blur these together, we lose clarity. The role gets diluted, and expectations misalign. If we want the profession to thrive, we must start with the role. What is it that a product manager is accountable for, empowered to decide, and measured by?
The other three — activities, knowledge, and operating model — exist to enable that role.
Product Management In The Age Of AI
And now, a new force is reshaping us again.
AI is changing how product managers think, create, and deliver. It’s giving product managers independence. We can analyse markets, test ideas, and communicate with breath-taking speed. We can spin up interactive prototypes, narratives, and data-driven insights in minutes. We can create working code without knowing how to code.
We’re becoming faster, sharper, and more prolific.
But we’re also becoming a little sloppy. Because when creation becomes effortless, debate becomes optional. AI tempts us to bypass the hard conversations or the friction that forges better thinking. It’s easy to generate options; it’s harder to choose the right one.
If we’re not careful, we’ll build the wrong thing faster than ever before.
AI can collate customer feedback, summarise interviews, and detect patterns we miss. But it can’t feel the frustration in a customer’s voice. It can’t read the pause in a conversation or the emotion in a moment. That’s still our work.
And it can’t navigate the organisational politics, the influence, the coalition-building that sits at the heart of our craft. Product managers still lead without authority. We persuade, align, and sometimes absorb the tension no one else will hold.
AI can make us productive. It can’t be persuasive on our behalf.
Balancing Speed With Depth
Across the product lifecycle, we have tools for almost every task, each one removing a layer of friction that previously required collaboration and debate. Here are the leading tool-types and exemplar products to consider, subject to fit with their context, capability, governance and adoption readiness.”
- Customer discovery: ChatGPT, Perplexity, Grain, Dovetail — summarise interviews, identify patterns, and generate customer insights in minutes.
- Market and competitor analysis: Knoema, Aomni, ChatGPT with browsing, Consensus — scan data, pull reports, and build comparative positioning instantly.
- Ideation and concept testing: Miro AI, FigJam AI, Notion AI — brainstorm features, generate problem statements, and create early prototypes.
- Design and prototyping: Uizard, Vercel v0, Lovable, Alloy — translate text prompts into mockups, flows, and production-ready front ends.
- Specification & documentation: ChatGPT, Notion AI, ChatPRD, ReqAIre — draft PRDs and feature specs, write epics/user stories and acceptance criteria, ensure compliance and traceability.
- Engineering and build: GitHub Copilot, Cursor, Devin — generate and debug code at speed.
- Testing and QA: Testim, Reflect AI, Qodo — create and run test cases autonomously.
- Launch and marketing: Jasper, Copy.ai, Runway, Synthesia — craft messaging, visuals, and videos in minutes.
We can research, ideate, design, and deliver all without needing a team. That’s exciting. And bloody dangerous.
Just because we can build it doesn’t mean we should.
Product management has always been about intentionality. It has always been about knowing why something should exist, not just how to make it real. AI doesn’t replace that judgement; it amplifies the consequences of getting it wrong. We can now make bigger mistakes, faster, and with far more polish.
Executives shouldn’t invest in product management to see more things built. They should invest in making better decisions. If we don’t slow down to think, to test the strategy, to pressure-test the insight, we’ll become a faster product factory, not a better one.
And it raises a deeper question: if AI lets us do everything ourselves — research, design, code, launch — are we still doing product management? Or are we quietly drifting into the role of maker, skipping the collaboration and debate that sharpen our thinking?
If we lose that friction, we lose the essence of the craft. Product management was never meant to be a solo pursuit. It’s a team sport, one built on challenge, on shared sense-making, and on the courage to pause when everyone else is racing ahead.
AI might let us move faster than ever. But the real skill is still knowing when not to.
What Product Management Really Is
To me, product management is the intersection of all four lenses: function, activities, knowledge, and operating model, which is anchored in the role.
We are growers of products and stewards of investment. We take fragile ideas and turn them into value. We create the conditions for growth. We understand the context, the positioning, and timing so that products can thrive.
We don’t just manage backlogs or roadmaps. We connect dots others can’t see. We translate complexity into clarity, and clarity into decisions.
In an AI-enabled world, that is our edge. The tools can make things. But they can’t align with our goals.
The Future Of Our Craft
No doubt, AI will keep changing how we work. But our purpose remains constant: to ensure the products we bring to life are worthwhile, intentional, and aligned with the organisation’s goals.
We are not operators of process.
We are custodians of purpose.
We are not just managing technology.
We are shaping humanity’s relationship with it.
That’s what product management is.
Not a process.
Not a role on an org chart.
But a practice of questioning, choosing, and shaping the future with intent.
So, what is product management going forward?
It’s the discipline of slowing down just enough to make sense of speed.
The courage to hold the long view when technology tempts us to sprint.
The practice of shaping the future, not just shipping it.
