In 2025 we'll see an increased adoption of agentic and other AI solutions to augment tasks in many areas of desk-based work, including product management. I’ll share my view on relevant areas of application in the appendix. While I don't believe that this will lead to PMs or product leaders being replaced it will certainly change how we work and what we should be focussing on. In simple terms we should happily leave machine-tasks to machines and focus on what makes us human. So far so good, but what does that mean for product leaders?

Emphasizing human qualities as a leader

I have always strongly believed in the importance of the more human aspects of leadership when working with smart individuals. I don’t see a contradiction between traits like empathy, authenticity, humbleness and vulnerability on one hand and assertiveness or performance orientation on the other. Earlier in my career as a product leader, this was not always recognized by senior managers who still promoted the ideal of a “bullish”, technocratic leadership style - which is why I was pleased to notice that apparently even the big management consultancies (that those leaders of mine were shaped by) meanwhile recognize the necessity of a more human approach to leadership for sustainable success. Check out this podcast with McK’s Senior Partner Dana Maor.

A human approach to leadership generally makes sense in dynamic times of high uncertainty and change - and as such would be as relevant even if we took AI completely out of the picture. On the other hand, the speed of change and the new opportunities & challenges that come with AI further emphasize the importance of showing your human side as a leader so that others will trust your leadership.

Specific focus areas in product leadership

In addition to this general attitude in leadership, I also see the need for some more product management-specific focus changes in how we lead in an AI-augmented work environment:

First, our role in coaching and developing PMs will need to evolve. Where previously we might have spent time helping PMs structure their user research or analyze competitive data, we'll need to shift towards helping them master the art of asking the right questions – both of their AI tools and their stakeholders. The challenge won't be in gathering insights (see Appendix: AI Tooling in Customer Understanding), but in knowing which insights matter and how to act on them.

This shifts our focus from teaching tactical skills to developing strategic judgment. How do we help PMs distinguish between patterns that matter and those that are just noise in their analytics? How do we ensure they maintain deep customer empathy when more of their interactions become AI-mediated? Product leaders will need to become masters at teaching the art of context-switching between AI-augmented analysis (see Appendix: Meeting Intelligence) and genuine human connection.

The way we structure and run our teams will likely change too. With AI handling more of the routine analysis and documentation (see Appendix: Product Specification & Planning), we might see product teams becoming more fluid, with PMs able to handle larger scope but requiring stronger collaboration skills. Our role will be less about ensuring execution quality and more about fostering the right collaborative dynamics and decision-making frameworks.

Here are key areas where I believe product leaders should focus their energy:

• Developing Judgment Frameworks: We need to create clear frameworks that help PMs understand when to rely on AI-generated insights and when to dig deeper through human interaction. This isn't just about decision-making matrices – it's about developing intuition for when the human touch is irreplaceable.

• Building Enhanced Feedback Loops: As AI accelerates the pace of product development, we need to establish stronger mechanisms for rapid learning from real human experiences. This means teaching PMs how to combine quantitative AI-driven insights with qualitative human feedback in meaningful ways. 

• Bringing Stakeholders Along: At the same time, the insights generated from a broader mix of sources - including previously opaque ones - require extra attention on communication that brings the rest of the org along. Several product organizations I know experienced pushback when shifting from “feature-request lists” in Salesforce to a structured collection of multi-source insights in tools like Productboard because commercial teams perceived it as a black-box approach. This effect will be even stronger with insights extracted and structured by AI tools, so we need to keep our commercial colleagues in the loop, for instance by regular updates about our learnings.

• Fostering Creative Tension: Our teams will need to become comfortable with the creative tension between AI efficiency and human creativity. Product leaders should actively create spaces where PMs can experiment with different balances of AI augmentation and human-led innovation.

• Cultural Navigation: As organizations integrate more AI tools, product leaders will need to become skilled at helping their teams navigate the cultural implications. This includes addressing fears about AI replacement, setting healthy boundaries for AI use, and maintaining team cohesion in an increasingly hybrid human-AI workflow.

Remember: A Fool with a Tool is still a Fool!

The assessment of PM performance will need to evolve as well. While it’s fair to be impressed by visible speed increases in generating insights, deriving potential solutions or shipping stuff we shouldn’t be misled to believe that faster is necessarily better - and we need to be careful not to be bluffed by smart tool usage.

If there's one clear danger in the increasing adoption of powerful AI tools in product management, it is that they may lead to solid-looking but, in essence, flawed work that is difficult to spot at first glance. A fool with a tool is still a fool—but me may not immediately notice. This is why it's crucial that product leaders constructively challenge their PMs' results to avoid accidental or intentional overreliance on AI enhancements.

New challenges for product leaders

This isn't about becoming AI experts – it's about becoming better at what humans uniquely bring to the table: nuanced judgment, emotional intelligence, and the ability to navigate ambiguity with wisdom rather than just data. The question for us as product leaders isn't just how to adapt to this change, but how to help our teams thrive in it.

Looking ahead, successful product leaders will be those who can create environments where both human creativity and AI capabilities can flourish. This means developing new skills ourselves: Becoming better coaches in an AI-augmented world, understanding the boundaries between human and machine decision-making, and maintaining our teams' humanity while embracing technological advancement.

It also means to accept and embrace change: With our teams as well as individually we better get ready to be pushed beyond the comfort zones of what we know and how we’ve been successful in the past. This can be uncomfortable, but if we don’t fall into the trap of seeing ourselves as victims of change and instead remain curious we have the opportunity to actively shape how human-AI collaboration will work in product development.

I’m curious to hear from you: What aspects of product leadership do you think will remain fundamentally human-centric? How are you preparing your teams for this shift? And most importantly, how do you envision the balance between human intuition and AI-driven insights in your product organization?

Appendix:
Examples for AI augmentation of Product Management work

All broader product management platforms like airfocus, Productboard or Pendo have meanwhile included AI based feature to analyze insights or generate user stories. But there’s also a range of usecase specific solutions:

Customer Understanding & Insights: Tools like Breyta, Dovetail or Next are turning unstructured qualitative data (such as recordings of CS calls) into a goldmine of product insight that previously would have taken weeks to compile.

Meeting Intelligence & Documentation: Notetaking assistants like Otter.ai and Jamie aren't just transcribing meetings anymore – they're extracting action items, creating summaries, and maintaining decision logs. Tools like Rewatch are making video meetings searchable and analyzable at scale.

Collaborative Work Processing: Solutions like Miro AI are transforming how we handle post-workshop synthesis, automatically structuring and categorizing workshop outputs into actionable formats.

Product Specification & Planning: Gluecharm or Userdoc change how we create and manage product specifications. These tools can expand high-level product concepts into detailed user stories and acceptance criteria. This allows teams to explore a broader solution space before committing to a direction. Other tools like Delibr specialize on collaboration around product requirements.

Data Analysis & Reporting: Tools like Amplitude's AI-powered analytics are automating the initial analysis of product usage data, flagging significant patterns and potential issues.

In addition to these usecases I can for instance also imagine AI augmenting solutions in areas such as compiling regular internal updates, comparing roadmap-scenarios.  

Please note that this is a non-exhaustive, subjective selection of examples in a fast moving market.

For those of you who read German and want to keep up with some of the upcoming new tools from a European perspective you might want to check out Das KI-Radar, a subjective collection of AI tools for product development by Alexej Antropov