AI is changing product work. Who governs the decisions?
Table of contents
Artificial intelligence (AI) is now part of almost every product (management) conversation. Teams use it to research, to design, to write, to test, and to decide. This is not a future scenario. It is already happening inside global organizations, across many countries and many teams at once.
Most of the public conversation about AI and UX still focuses on tools. New features, new plugins, new ways to generate a mockup in seconds. This is useful, but it misses the harder question. The real challenge is not the tool. It is governance.
Governance is the first problem, not the last
When an AI model helps write a research summary, suggests a design pattern, or scores user feedback, someone still has to decide what to trust and what to reject. In a small team, this can happen informally. In a global product organization, it cannot.
The scale of the problem is not small. The global market for AI governance tools and frameworks is expected to grow more than tenfold over the next decade, which is a good signal of how far behind most organizations still are. And the regulatory backdrop is moving fast: the EU AI Act reaches full effect in August 2026, alongside frameworks like the US NIST AI RMF and ISO/IEC 42001, so organizations operating across several countries now have to reconcile more than one rulebook at once.
Complex organizations need clear answers to simple questions. Who owns the decision when an AI tool proposes a design direction? Who is accountable if an AI-assisted research summary misses something important? What level of human review is required before an AI output reaches a customer? One recent analysis of product management put it well: governance in 2026 is not a compliance checkpoint to clear, it is the mechanism that decides which initiatives get funded and which get stopped. That is a very different posture than treating governance as paperwork.
These are not technical questions. They are governance questions, and in practice they have quietly moved from being legal’s problem to sitting directly in the product manager’s backlog. Right now, most organizations do not have clear answers.
Management: the skill gap is not about prompting
A lot of AI training today focuses on prompting skills, on how to ask the tool the right question. This is a small part of the real problem.
The bigger gap is management. Product and UX managers now need to manage mixed teams, where some of the work is done by people and some is done by AI systems. A simple rule is emerging across teams that are getting this right: whoever uses an AI output is accountable for vetting it, and AI-generated work gets the same quality gates as human work, not a lighter version of them. That sounds obvious, but very few teams have written it down as a standard.
This changes what a manager actually reviews. As one 2026 field guide for product managers describes it, the skill that now separates effective AI-augmented managers from careless ones is spotting what an AI draft missed, oversimplified, or distorted, not producing the first draft faster. And teams that take this seriously are starting to formalize it: structured evaluation practices, sometimes called evals, are becoming a product responsibility rather than only an engineering one, because someone has to own the definition of “good enough” before it reaches a user.
This is a new kind of management skill. It sits between traditional people management and traditional quality assurance, and very few managers have been trained for it.
Leadership: setting principles before the standards exist
Global, complex organizations rarely have the luxury of waiting for industry standards to mature. Leaders have to set internal principles now, with incomplete information, and be ready to adjust them as the field moves.
Figma’s 2026 AI report, based on more than eight thousand survey responses, found something worth sitting with: organizations where AI adoption is either purely top-down or purely grassroots run into friction, because leadership strategy and practitioner reality are disconnected. The goal the report describes is “unified” adoption, where leadership direction and daily practice reinforce each other instead of pulling apart. That is a leadership design problem, not a tooling problem.
Some UX and research leaders are already documenting how they approach this. One research leader described setting an explicit north star for how AI should be integrated into the practice before rolling any tool out to the team, precisely because how a leader frames AI sets the tone for how the whole team will use it. Another recent piece argued that design leadership in the AI era is primarily change management, because adopting new tools means retooling how people spend their day, not just what software sits on their desktop.
This means leaders need to decide, today, what “responsible use of AI in product work” looks like for their own organization. Not as a policy document nobody reads, but as a working set of principles that guides daily decisions: what data can go into an AI tool, what level of transparency is owed to users, what human oversight is non-negotiable.
Leadership in this context is less about knowing the answer and more about building the structure that lets good decisions happen consistently, across teams and across countries.
Why this matters for people who hold product decisions
This is exactly the space UX-PM was built for. The certification does not train people to be better designers only. It trains the people who make product decisions, wherever they sit in the organization: product managers, UX leads, researchers, and the people above them who set direction.
The AI conversation inside product work is, at its core, a governance, management, and leadership conversation. It asks the same question UX-PM has always asked. Who is responsible for the decisions that shape the product experience, and are they equipped to make them well?
This is the first article in a new series where we open that conversation. In the next articles, we will go deeper into each of these three areas: governance frameworks for AI-assisted product work, the management practices that hold mixed human-AI teams together, and the leadership principles that global organizations are starting to adopt.
Further reading
- Product management in 2026: from experimentation to the reckoning (Metis Strategy)
- How AI Is Changing Product Management in 2026 (Field Guide)
- AI in Product Management Guide for 2026 (Gocious)
- How To Build An AI-Native UX Team Without Losing Design Quality (citing Figma’s 2026 AI Report)
- Thoughtful AI implementation for UXR leaders (UX Collective)
- In the age of AI, the UX field survives on leaders who cultivate juniors (UX Collective)
- AI Governance and Regulation 2026: A Complete Guide to Global Frameworks
- Product team challenges in 2026 (Troy McAlpin)
- How AI is Changing the Product Manager Role in 2026 (Codebasics)
- AI evals for product managers: the complete guide for 2026 (Lovelaice)