Brainstorm and rank risks for a project plan
Feed in a project plan and get risks ranked by likelihood times impact, each with an early-warning sign and a concrete mitigation to plan around.
Last updated
Fill in the blanks
Assembled in your browser — nothing you type is stored or sent anywhere.
Your prompt
Model-agnostic — works in Claude, ChatGPT, Gemini or any AI chat.
Why this prompt works
- →Requiring likelihood, impact and a sort order forces a ranked list instead of a flat brain-dump, so the reader's attention goes to the risk that actually matters most rather than whichever one the model wrote first.
- →Demanding an early-warning sign per risk, not just a mitigation, is what makes the assessment usable during the project instead of only useful in hindsight after the risk has already landed.
- →Rejecting "monitor closely" as a mitigation and requiring the risk to trace back to something in the plan keeps the model from padding the list with generic risk-management filler that isn't grounded in this specific project.
Part by part
- an early-warning sign someone could actually notice
- The clause that makes a risk register operational. A risk with no observable precursor cannot be watched for, only regretted.
- a specific mitigation
- Names the exact filler this exercise attracts. The prompt spells out what does not count, because "monitor closely" is what a register says when nobody has decided what to do
- Sort the list with the highest likelihood-times-impact risks first
- Ordering by the product rather than by either factor alone, so a near-certain annoyance does not outrank an unlikely catastrophe.
- label it as speculative rather than presenting it as certain
- Separates risks the plan supports from ones the model inferred. Mixing the two makes the whole register harder to trust.
Common mistakes
What are the risks with this project?Returns generic project risks — scope creep, key person dependency, underestimated timeline — that apply to every project ever run. True, unranked and unactionable.
List risks and mitigations for this plan.Without the early-warning requirement, mitigations arrive as intentions rather than triggers. A mitigation you cannot tell when to apply is a paragraph, not a plan.
Identify all possible risks.Exhaustiveness produces a list too long to act on, where the serious risks are diluted by unlikely ones. Ranking is what makes a register something people revisit.
Questions
Why insist on an early-warning sign?
Because it converts a risk into something you can watch. "The vendor may be late" is a worry; "no test environment by the 12th" is a date on a calendar that someone can check.
What is wrong with monitor closely?
It commits nobody to anything and cannot be evaluated afterwards. A mitigation should name an action, an owner and a trigger, or it is a way of writing a risk down without addressing it.
Why multiply likelihood by impact?
Because either alone misleads. A high-likelihood, low-impact risk sorts above a low-likelihood catastrophe on frequency, and below it on severity — the product is a crude combination and better than either.
Should speculative risks be included at all?
Yes, labelled. Some of the most serious risks are not visible in the plan, and excluding them narrows the exercise to what was already written down. The label is what keeps the register honest about its evidence.