Risk brainstorming and pre-mortems
Use the model as an extra participant in risk analysis: have it imagine the feature has failed in production and describe how, then rank the failures for testing.
How it works
A pre-mortem asks a group to assume a project has already failed and explain why; it surfaces risks people hesitate to raise directly. The model takes part as one more voice: given the feature, its users and its context, it writes plausible failure stories — the data migration that silently truncated a field, the integration that timed out under month-end load, the permission change that exposed one customer's records to another. Each story is then turned into a risk with an estimated likelihood and impact, and the ranked list shapes where testing effort goes. The model is also useful afterwards, to check a finished risk list for categories it doesn't touch at all.
When to use it
At the start of a feature or release, before test planning; for high-stakes changes such as payments, migrations and permission models; when the team is too close to a feature to see its weak points; and for solo testers who have nobody to run a risk session with.
Limitations
The model's failure stories draw on common patterns, so it is good at familiar risks and weak at the ones specific to this system, organisation or market. Likelihood and impact estimates from the model are guesses dressed as numbers; it has no data about your production. A long list of plausible risks can create the impression of thoroughness while missing the one risk that matters.
Watch for
The team edits the model’s risk list instead of writing its own, so the risk only someone in the room knew about, such as a batch job sharing the same table, never gets said. Have people write their own risks before they see the model’s, then merge the two lists.
What you still own
The final ranking, and the call on which risks testing will address and which are accepted or handed elsewhere.
Cite this
Qlarify Labs. (2026). Risk brainstorming and pre-mortems. Retrieved from https://labs.qlarify.fi/ai-in-testing/risk-brainstorming-and-pre-mortems