FRAME-07ExecuteSave

Risk register and RAID log

A running list of what could go wrong, how likely and how bad it is, who owns it and what you're doing about it, extended in a RAID log to assumptions, issues and dependencies.

Get the template

TPL-04 · v1 · updated Nov 17, 2026
BlankRAID log with dropdowns, an automatic Red/Amber/Green rating and a likelihood × impact grid.Excel .xlsx · 13 KB
Filled exampleFilled in for the running example: five teams moving a support assistant onto in-house inference.Excel .xlsx · 14 KB

Free to use and adapt. Save this page to get an email when the template is updated.

When to use it

Start the register at kickoff and keep it open until the retro. A risk is something that might happen; an issue is something that already has. The register exists so risks get an owner while they are still cheap to deal with.

Risk register vs RAID log
UseWhen
Risk registerMost programs: you mainly need to track what could go wrong and who owns it
RAID logPrograms with many teams or vendors, where assumptions and dependencies cause as many surprises as risks
NeitherA short, single-team project: a risks section in the plan is enough

A filled example

The top risks on Lantern, plotted by likelihood and impact. Anything in a red cell gets reviewed every week.

Lantern risks by likelihood and impact
LikelihoodImpact: LowImpact: MediumImpact: High
Likelihood: HighAmberRedRed
  • R-01 Index rebuild too slow for the cutover window
Likelihood: MediumGreenAmber
  • R-03 GPU capacity too tight at peak support traffic
Red
  • R-02 Security & privacy review not scheduled in time
Likelihood: LowGreenGreen
  • R-04 Only engineer who knows the serving config on leave in launch week
Amber

And a few rows from the RAID log, one of each type:

Sample RAID log rows for Lantern
TypeIDDescriptionOwnerStatus
RiskR‑01Index rebuild too slow for the cutover window; mitigation: run it in parallel batchesData leadOpen
AssumptionA‑01The vendor API stays available for rollback after cutover, until the contract endsEng leadConfirmed
IssueI‑01Staging GPU cluster down, blocking the second shadow runEng leadBlocked
DependencyD‑01Security & privacy sign-off before customer traffic movesSecurityOpen

Good vs bad

Common risk register mistakes and how to fix them
FlagWhat it looks likeFix
No ownerA risk owned by “the team” or by nobodyOne named person per risk
No mitigationA risk written down and never acted onEvery red or amber risk has a next action and a date
Vague risk“Timeline risk” with no cause or effectWrite it as cause, event, effect: “Because X, Y may happen, which would Z”
Never closedA list that only grows, so nobody reads itClose or downgrade risks at each review; archive closed ones

How to run it with a team

  1. Seed the register at kickoff: ask each workstream lead for their top worries and write them as cause, event, effect.
  2. Rate likelihood and impact together, so the scale means the same thing to everyone.
  3. Give every risk one owner and a next action, then agree which ratings get reviewed weekly.
  4. Review the red and amber risks in the weekly program meeting; skim the rest at milestones.
  5. Pull the risks that changed into the weekly status report.

Inside the template

Template fields and what to put in them
FieldWhat to put there
ID and typeR, A, I or D plus a number, so items can be referenced in status reports
DescriptionCause, event and effect in one sentence
Likelihood and impactLow, Medium or High each; the rating cell works out the RAG color
OwnerOne named person
Mitigation and next actionWhat you are doing about it, and by when
StatusOpen, blocked or closed; assumptions are confirmed or not

Get the template

TPL-04 · v1 · updated Nov 17, 2026
BlankRAID log with dropdowns, an automatic Red/Amber/Green rating and a likelihood × impact grid.Excel .xlsx · 13 KB
Filled exampleFilled in for the running example: five teams moving a support assistant onto in-house inference.Excel .xlsx · 14 KB

Free to use and adapt. Save this page to get an email when the template is updated.