TL;DR. AI now drafts the requirements, builds the prototypes, summarises the user feedback, and writes production-ready software from little more than a product brief. It does this in days, and the quality of the deliverables rises every month. Those artefacts used to reflect the real work: a polished document would stand in for the thinking behind it. So what is left for product managers, now that AI performs the most time-consuming parts of the role? Bandwidth. The real opportunity I see for ambitious product managers is to run several initiatives in parallel, experiment faster, reach the market sooner, and make faster decisions backed by data and market evidence. That shifts the role significantly: a more demanding job, heavy on experimentation and go-to-market, where time-to-market, rapid experimentation and fast decision-making are the success factors. Four qualities make that job possible: entrepreneurial thinking, product and user sense, cross-domain knowledge and a strong technical background. AI makes each of them more important, not less: Product managers at innovation mode.
The Artefacts Were Never the Job, but a Major Part of It
A large part of a product manager's week used to go on producing artefacts: requirements documents, user stories, wireframes, competitive summaries, feedback reports, release notes. In 2026, AI produces good first drafts of all of them, and it no longer stops at drafts. Given a clear brief, agentic coding tools deliver working, tested software. A product manager can describe a concept in the morning and put a working prototype in front of users in the afternoon, and a small product can go from vision to first release in days.
Three consequences follow. First, high-quality, sophisticated outputs no longer carry the significance or the signal of deep thinking they used to: anyone can produce a polished PRD, and increasingly anyone can produce a polished product. In this sense, whether it is the right product matters more than ever: success for product managers in the AI era is about the judgement on where to focus, and how to go to market and grow fast. It is about building a product innovation and experimentation machine: feedback loops that took a quarter close in days, a draft concept becomes a working experiment in hours, users react to it the same week, and the next iteration ships while the last one is still being measured. Product innovation runs at the pace of experimentation now, and every product call is scored while everyone still remembers who made it.
Product innovation is no longer a side initiative; it runs continually, at the core of product development.
Second, the weight of success moves to making it happen, where making it happen is about taking, growing and establishing the product in the market. It was never only about building or shipping, but now that those are the easy parts, it is, more than ever before, about making the product truly viable in the real world.
Third, now that the artefacts are easy, people from other domains increasingly act as product managers. That can be good, more minds thinking in products, but it also creates noise and misalignment, and resolving that lands on the product manager. Personally, considering the massive impact AI is about to have on software developers, I expect to see plenty of former software engineers excelling in product leadership roles: it is a meaningful career shift in many ways, and they arrive with the technical depth the role demands. And product managers can lead this broader role transformation and fusion.
One warning. The same tools that power the experimentation machine also make it cheap to look productive: ten plausible prototypes a week look like momentum whether or not any of them matters. Companies that measure outcomes will tell the difference; companies that admire artefacts will be fooled faster than ever.
So What Is a Good PM in 2026?
Bandwidth is only worth what you spend it on. A good PM in 2026 spends it on what I have always preferred to call product leadership rather than product management.
They make things happen by releasing the right instances of the product, in the right order. They connect the technology and commercial worlds to create value for users and for the business, and they act as the CEO of the product.
Product leadership is the future of product management, and it is the AI-proof part of the role: machines now write the documents, and increasingly the software itself, but setting the direction and answering for the outcome is not something you can prompt for. The centre of gravity has moved: fewer artefacts, more decisions. Smarter decisions, grounded in data rather than opinion; faster decisions, because the evidence arrives in days and the market does not wait; and, after the decisions, the market work: reaching it sooner, and testing, refining, validating and establishing the product there. The role concentrates into the four qualities that always decided it. I described the full set of traits in Product Leadership: 9 Traits of Great Product Leaders; here is what each of the four becomes when AI does the production work.
1. Entrepreneurial Thinking
This is the foundation. Product leaders are enthusiastic about designing and building products people love. They are driven by innovation, and they stay connected with reality through clear objectives and systematic measurement of performance and success. They think big, and they know the constraints and priorities: how to use the available resources for the right instance of the product.
What changes with AI: the cost of building has collapsed, so the cost of building the wrong thing is mostly time and attention. Entrepreneurial thinking now means saying no more often, and choosing the few bets worth a team's focus. It also means asking what the product owns. If a competitor could rebuild your product with the same model API in a weekend, the advantage has to come from somewhere else: data, distribution, workflow or trust.
2. Product and User Sense
Thinking in products is not as easy as it sounds. Over the years I have watched promising ideas forgotten or misunderstood, and teams struggling with the very notion of a product and how to develop one. A product leader sees the whole product: what it is, how it can be built and how it will succeed, and the market around it: competitors, potential partners, the technology, the differentiators and the possible monetisation models.
You need to be able to start with an idea, derive a concept and define a product with clarity. And you need to think as a user, with the empathy to relate to their problems, frequent needs and expectations, and to define success in measurable terms.
What changes with AI: AI can summarise what users said; it cannot tell you what they meant, or what they will do. Product sense now includes knowing when an AI summary of feedback is flattening the one signal that matters. For AI-powered products it also needs a new kind of user sense: understanding how people experience a product that is right most of the time rather than all of the time. I wrote about that shift in From Deterministic to Probabilistic: Rethinking Product Requirements for the AI Era.
3. Cross-Domain Knowledge
Modern digital products are complicated. Success depends on technology, usability, business models, marketing and operations working together. Product managers work with designers, engineers, data scientists, marketing and commercial teams, and must understand the language, perspective and capabilities of each.
For a consumer product, for example, a good product manager understands the principles of user experience, how engagement is measured and the enabling technologies, interprets insights as part of continuous improvement, and knows when to run experiments and user tests. They spot the chance to deliver a better experience powered by back-end technology or an analytical model.
What changes with AI: AI is now one of the domains, and arguably the one that shapes all the others. Product managers need enough understanding of models, evaluation, data and cost to hold a real conversation with the people building them: what quality the product needs, how it will be measured, what happens when the model is wrong, and what each answer costs to produce.
4. A Strong Technical Background
In a digital context, a strong technical background is a must. Defining outstanding products depends on understanding both the capabilities and the constraints of technology. You can delegate technical work and decisions, but a solid grasp of the state of the art, global technology trends and the competition is a major advantage.
The opposite risk is real too. Too much focus on technology, especially for people from a pure engineering background, can cost you focus. Technical excellence is one part of a product's success, to be balanced with the commercial and operational parts.
What changes with AI: a technical background now includes hands-on fluency with AI tools. A product manager who can prototype a concept with an AI coding assistant, or test a prompt chain with five users before any engineering starts, learns in days what used to take a quarter. That fluency is quickly becoming the baseline rather than the differentiator.
What the Role Gains
The era does not only take work away from the role; it adds work that did not exist. When software is generated rather than hand-built, the quality bar becomes the product manager's main artefact: the evaluation sets, the examples of right and wrong answers, the definition of good enough that decides whether a build ships. When every answer a product gives has a cost, unit economics become a design decision rather than a finance report. And when products behave probabilistically, someone has to decide where 'right most of the time' is acceptable and where it never is. That someone is the product manager.
When building is cheap, the quality bar becomes the product manager's main artefact.
The Rest Still Matters
Commercial acumen, the ability to execute, an innovation mindset, the judgement to choose the right framework for each decision, and a leadership style that inspires rather than controls all remain part of the role. They are covered in the nine traits. For the team around the product manager, see How to Set Up and Lead a Great Product Development Team.
Great products are built by great teams, inspired by passionate, skilful product leaders.
If you are defining AI-powered products, the Product Toolkit gives you a structure for requirements in the AI era.
References
Krasadakis, G. (2018). How to become a Great Product Manager. Medium. The 2018 original this article reframes for the era of AI.
Krasadakis, G. (2026). Innovation Mode 2.0: Designing Innovative Companies in the Era of Artificial Intelligence. Springer, Cham. https://doi.org/10.1007/978-3-032-00835-0
