When building is much faster, is product discovery still important? At this September's session at Indonesia Product Conference 2026 (IPC 2026), one answer from the room changed the question. If you can build quickly, building itself becomes a way to learn, and the line between discovery and delivery gets harder to draw. What remains is knowing when discovery is still worth doing, and how to make the case for it when stakeholders want to skip it.
We explored this in an open roundtable, covering when it is acceptable to skip discovery, how discovery adapts with AI, and how to make the case for it.
"When building is fast, building can itself be part of discovery."
What we took away
Deciding whether to skip discovery
- •Discovery is often skipped when stakeholders have already decided on a solution, or when competitors already have a feature the team needs to match. Sometimes matching competitors is the right decision
- •The smaller the expected impact, the more acceptable it is to skip discovery. The larger the impact, the more discovery is worth doing
- •If the problem is well known, teams can move straight to building. If the problem is new, discovery is worth doing
- •The longer something takes to build, the more discovery is worth doing first
How discovery adapts with AI
- •When building is fast, building can itself be part of discovery. Teams can build something quickly and learn from it, so discovery and delivery overlap more than before
- •AI makes it possible to go straight to high-fidelity designs, but low fidelity still has a purpose. It keeps the discussion on the problem, while high-fidelity designs tend to pull feedback towards colours, illustrations and icons
- •Whether AI-generated users can replace talking to real people depends on how accurately AI can predict human behaviour
Making the case for discovery
- •When stakeholders ask to skip discovery, start with the data you already have, rather than pushing back straight away
- •Check whether that data supports their proposed solution. If it shows that important information is missing, use that to argue for more discovery
- •Some of that missing information can only come from talking to users directly, because teams often do not know what they are missing until they do