How Agentic AI will redefine the role of the Product Manager

Every few years, tech throws up a question that makes Product Managers shift uneasily in their seats. This is one of those you may have heard recently: will agentic AI—those autonomous, decision-making digital workhorses—render SaaS platforms obsolete? And if that happens, what happens to the Product Managers who build them?

Doomsayers have been suggesting for a while that Product Management might be near its end, at least since Airbnb got rid of the PM function, but most likely the job is just transforming. If you’ve spent years honing your skills in managing feature roadmaps, refining user journeys, and optimising conversion funnels, you’ll still have a role. But if you ignore AI’s impact on how software is built, sold, and used, you might find yourself managing a product no one needs.

Remember when tech killed everything and then it didn’t?

Tech has a habit of reinventing itself without wiping the slate clean. We’ve been here before, more times than we care to admit. Remember when mobile apps were going to kill the web? When cloud was supposed to kill on-prem? When no-code was going to kill developers? None of these things happened. Instead, we got a messy, layered ecosystem where the old and the new coexisted, adapted, and found their niche.

Agentic AI, for all its promise, won’t make SaaS disappear. What it will do is shift the value proposition of SaaS products. SaaS has always been about providing structured workflows, prebuilt logic, and intuitive interfaces. AI agents, on the other hand, promise something different: decision-making autonomy, cross-platform orchestration, and outcome-driven automation. The two aren’t mutually exclusive—at least, not yet.

What this means for Product Managers

If agentic AI changes the nature of software, it also changes the nature of Product Management. The old playbook—customer research, problem identification, feature prioritisation—still matters. But AI-first products demand a new layer of thinking. Here’s what’s shifting:

From feature-centric to outcome-centric thinking

This should have been the reality a long time ago, but too many PMs still think in terms of features and outputs. In SaaS products, and B2B SaaS especially, this is still visible. AI-first products are strictly about outcomes. The question isn’t “How do we help users complete this task?” but “How do we get the task done with minimal input?” Product Managers will need to redefine success metrics, moving from usage-based engagement to goal-driven automation.

Less UI, more API

If AI agents handle tasks autonomously, the traditional UI-heavy approach may become less relevant. Instead, the focus shifts to API design, integration depth, and how well your system plays with others. That also means trust becomes a key product attribute—users need confidence that AI-driven decisions align with their goals and ethics. Transparency, control, and feedback loops become part of the product.

Data as a moat, not just a feature

AI doesn’t run on code alone; it runs on data. While SaaS companies have long understood the importance of data, AI shifts the stakes. The winners won’t just be those with the best algorithms but those with high-quality, domain-specific data. Product Managers will need to think about data strategy as a core part of their role—where it comes from, how it’s structured, and how it contributes to competitive advantage.

The art of human-AI collaboration

For all its power, AI isn’t perfect yet and won’t be for a long time. It makes mistakes, lacks context, and sometimes does exactly what you asked—but not what you wanted. This means Product Managers must master the art of human-AI collaboration. When should AI take over, and when should humans intervene? What does an escalation path look like? Designing for explainability and fallback mechanisms will become just as important as designing for efficiency.

Disrupt or be disrupted

SaaS companies face a choice: embed AI deeply within their products or risk being bypassed by agentic systems that no longer need their interfaces. Product Managers will play a central role in this transition. The hardest part? Pushing their own companies to embrace AI-driven disruption rather than clinging to existing business models. We’ve seen this before—think of software vendors resisting the shift to cloud or media companies underestimating digital-first competitors. The lesson? It’s better to disrupt yourself than to be disrupted.

So, what’s next?

For Product Managers, now’s not the time to panic—it’s time to evolve. The fundamentals of Product thinking—understanding users, defining problems, and delivering value—remain unchanged. But how we execute will look very different in an AI-first world:

• Learn the language of AI: We don’t need to become a machine learning engineer, but understanding basic concepts will set us apart.
• Get comfortable with uncertainty: AI doesn’t always behave predictably. Learning how to manage probabilistic systems, rather than deterministic ones, is a new skillset.
• Think like a systems designer: AI agents don’t exist in isolation. They interact with multiple platforms, data sources, and human inputs. Designing for these interactions is key.
• Advocate for AI ethics and governance: The best AI Product Managers won’t just optimise for efficiency. They’ll push for transparency, fairness, and accountability.

The rise of agentic AI doesn’t spell the end of SaaS, and it certainly doesn’t mean the end of Product Management. But it does mean the job will look different. Product Managers who embrace AI, rethink their approach, and push their companies to adapt will still find themselves in demand.