2025 didn’t just accelerate execution. It raised expectations. As AI took over the “toil” and quietly threaded itself through every product workflow, something surprising happened: the fundamentals of product management became more important, not less.
Hosted by Jarom Chung, VP of Product at Lucid, a panel including Malte Scholz, Head of Product at airfocus by Lucid; Lauren Kearney, Director of Health & Platform Products at Twitch; and Chris Butler, Director of Product Operations at GitHub echoed the same theme: AI accelerates output, but it puts pressure on PMs to elevate their input.
“2025 was the year where AI ate the easy stuff, and then really heavily collapsed on the hard stuff,” said Malte Scholz, Head of Product at airfocus by Lucid.
With the barrier to execution collapsing and AI agents effectively “joining the workforce,” as Jarom said, the real differentiators aren’t velocity or the clever use of tools. They’re judgment, storytelling, prioritization, and the taste to uphold quality when speed makes it easy not to.
“We saw a really strong improvement in chat agents addressing unambiguous challenges that our users had,” said Lauren Kearney at Twitch. “That opened up human thinking and human judgment for more ambiguous cases.” Her team also saw an acceleration in doc creation. “These writing agents are allowing us to not only generate artifacts that align other people, but allow us to have a co-thinker, where we can brainstorm and iterate quickly on our own thoughts.”
“Everybody started using AI,” agreed Chris, “especially generative models, to be able to do things like doc generation: summarisations, and stuff like that. But I also think that people didn’t really tell anybody.” Why? Stigma, and trust.
So does winning in 2026 come from vibe coding, more prototypes or generating artifacts faster? Or will come from the human skills that AI cannot replicate: deciding what matters, communicating it clearly, and raising the standard of what teams deliver, especially now that they can deliver much more, much faster.
Here’s what the panel decided.
1. The new PM currency: Trust, transparency, and taste
In 2025, AI quietly threaded itself through nearly every product workflow. And one of the impacts is that trust, not tooling, has become the real differentiator for product managers. The speed at which PMs can produce artifacts no longer impresses teams, nor is it the marker of success. Instead, the scrutiny is shifting to how thoughtfully those artifacts are created.
“When I see certain AI documentation or an email written by AI, and it's clearly not proofread, credibility all of a sudden plummets. So, maybe the question is, how do we maintain credibility while leveraging these tools?” asks Jarom Chung, VP of Product at Lucid.
Rather than hide AI’s fingerprints, the panel believes it’s better to surface them. “Being upfront is more ethical,” said Chris Butler of GitHub. “Disclose the conversations and sources you used with an LLM. I think it is more ethical in some way.” AI-assisted work isn’t about the tool itself, it’s about whether the PM demonstrates ownership over the final product.
For Lauren Kearney of Twitch, the principle predates AI entirely. “I agree, and I don’t think that’s a new recommendation. That’s an evergreen writing skill: put your sources,” she said. “The dust will settle, but I think as we go through that transition, being ethical and being transparent about how you are working is a really important piece to upholding your credibility among a team.”
2. Product sense, product taste, and vibe coding: The critical balance
“Product sense, to me, is the ability to understand the problem, the user, the constraints, and the broader context around the product,” explained Malte. “Then, to really make decisions that actually generate an outcome. It requires judgment under uncertainty and risk. Product taste, to me, is something around feelings that comes typically from experience.”
Where is the balance?
“Should product people be more technical or more lean towards this judgment? It’s a tough one. “I like that product managers can go out there, create these prototypes in a much faster way,” said Malte. “On the other hand, this is a critical time that they won’t spend on the problem space, the discovery, the critical thinking, the alignment work they have to do.”
But there is, Lauren challenged, a difference between a technical product manager and one who’s vibe coding prototypes.
“If you want to use vibe coding to prototype things, okay. But I don’t think that’s what makes a technical product manager,” said Lauren. “The way that I define a technical product manager is you understand the technology and systems that are used to build the outcomes that you need for customers, and you can sit in a room of engineers and smell when something smells funky.
“Some of the times I angered my engineers the most was when I put on my technical hat,” agreed Chris. “My job is actually to center us on the business needs and the customer.”
The conclusion is that PMs should be prototyping. The problem occurs when leadership expects that PMs will vibe code something, and then it will go into production. “That’s the biggest way that every vibe-coded startup fails… with major security and architectural concerns,” said Chris.
3. What matters more in 2026: Storytelling, prioritization & quality
As AI accelerates execution, the real leverage shifts back to human judgment: what PMs choose to focus on, how they communicate it, and how they guide their teams through ambiguity.
“A PM has two jobs: storytelling and prioritization,” said Lauren. With engineering velocity increasing, PMs can no longer justify micromanaging tasks or obsessing over implementation details. Their value lies in shaping narratives that align teams and making the tough calls about what matters most.
But that requires discipline, and a willingness to abandon the busywork that often masquerades as productivity. “Product managers like to go where the pain is not. They love ‘busy work’,” said Malte. “It was perfect PRDs, in 2026, it might be vibe-coding prototypes. But we need them thinking about the big questions.” The temptation to hide in tools, templates, and toy prototypes is real, but the future rewards PMs who spend their time understanding markets, customers, and strategy rather than polishing artifacts.
Ultimately, the skill that may matter most is navigating ambiguity with confidence. “We have to be great at wrangling uncertainty, not solving it,” said Chris Butler of GitHub. “Help the team understand uncertainties so they can make better decisions today.” AI can accelerate workflows, surface patterns, and compress execution cycles, but only humans can make sense of what’s still unknown.
If 2026 has a mandate for PMs, it's this: Elevate the work. Focus on clarity, intent, and quality over volume. Let AI handle the churn, while product managers double down on the judgment and storytelling that drive real impact.
4. AI will speed up execution, but it may also increase toil
As AI becomes a ubiquitous part of product development, teams are discovering a paradox: while AI can dramatically accelerate execution, it can just as easily create new layers of work for product managers. Speed, it turns out, doesn’t always translate into ease.
For Chris Butler, the biggest breakthrough in 2025 wasn’t raw automation, it was perspective. “One of the most beneficial things was using opinionated, biased copilots to get feedback at every stage,” he said. “It helps PMs stop polishing documents forever out of reputational fear.” These AI copilots give you a low-stakes way to pressure-test ideas early, reducing the time spent endlessly refining artifacts before sharing them with real humans.
But not all acceleration produces relief. “It’s really true for startups that we’ll be able to explore more moonshots more quickly,” said Lauren. “But I don't think velocity is going to increase. I think toil will increase because I think people are going to all think that their idea is the best idea.” When everyone can prototype in minutes and ship experiments faster, the volume of work and the noise around what should be prioritized grow. Instead of narrowing focus, AI risks flooding teams with even more partially formed ideas, half-built features, and “just one more” experiments to look at.
5. Research is changing, but human insight still wins
As AI reshapes nearly every corner of product development, user research is entering a transformative, but precarious, new phase. Tools that can simulate personas, generate behavioral patterns, or even run synthetic usability tests are becoming increasingly accessible. Yet the panelists agreed: these innovations are useful accelerants, not replacements.
“I think how product managers do research will change,” reflected Lauren. “ There’s a lot of opportunity around having agentic users with different personas to test a product for you. Will this happen in 2026 in a real way? Maybe if somebody gets lucky. But I do think it’s something that is in the future. But does that change what product management does? No. You still have to use research. The question is, what inputs are you getting, and from whom? Would I ship something based only on agent feedback? I wouldn’t.”
“I do think the problem that I've seen for product managers is that they are not as good at user research as they think they are,” said Chris. “There are lots of PMs that I've met that just want to have a conversation with you about a product, but it's like a sales pitch. That's not user research. There is a real value in using LLMs to coach and train PMs on asking better questions of customers.”
If we use LLMs to help PMs ask better questions, that’s when user research gets better, not replaced.
“I think it's a great point,” agreed Lauren. “I think we read all these articles saying that Generative AI is going to be able to do 90% of all jobs in the world. But that does not translate to 90% of all jobs are going to be replaced. Just because a tool can do a thing doesn't mean that it should, or that it does it well. It just takes away some of the toil.”
6. 2026 requires vulnerability, coaching & boundaries
As the conversation turned from tools to people, a quieter theme emerged: the future of product work will demand more vulnerability, not less.
“I think it comes down to making yourself vulnerable,” said Malte. “The job has gotten much harder in all dimensions. In a way, the need for guidance and coaching is heavily increasing, and many people don’t want to admit it. My recommendation always to my direct reports has always been, please please please reach out when you don’t know, because I don’t know all the time, and I think there needs to be more interpersonal coaching.”
Chris built on that idea by reframing where learning actually happens. It’s not just in frameworks or postmortems, but in the space teams intentionally create to think together. “Make space for the learning you need to do,” said Chris. “That’s the learning from your team about what you’re doing, learning from your leaders, and how they make decisions. I think it’s really important: providing the space for better decision-making and discussion, discourse. We don’t want every meeting to just be another status meeting.”
Chris argued for using AI to strip away the repetitive toil of status updates and coordination, but warned against outsourcing the human work of shaping meaning. “Find those spaces and then find the automations that can help you take away some of that toil that is usually around communication flows, like status updating. But also don’t trade off the human shaping and signaling of the messaging.
Too often, PMs burn cycles translating information into narratives they think leaders want to hear. “The PM should be adding information, and the leader should be interpreting it and using tools to interpret it,” he said. When the discussion circled back to what grounds product work amid all this change, Lauren offered a simple anchor. “When in doubt, be the voice of the customer,” advised Lauren. “Because, as intelligent or intuitive or accurate as Generative AI may or may not get, it will never be human. Technology will never have the human experience, and it is the human population that is using the products. And so, when in doubt, if you don't know what to do, or you don't know what the technology is, or you're burnt out from trying to learn every single tool that your, you know, leader mentions in a week and maybe there's a dozen of them, just recenter around being the voice of the customer in any room, in any conversation. It is a valuable mechanism to recenter your team around what really matters.”
https://www.youtube.com/watch?v=mTxmuBm9mYo&feature=youtu.be
Emma-Lily Pendleton
Read also
Experience the new way of doing product management
Experience the new way of doing product management