See risk differently. Decide with perspective.
Most tools ask how big a risk is. Perilometry also shows how differently the people deciding see it, and why.
See it.
Everyone at the table already holds a picture of the risk. Perilometry shows every view side by side.
Understand it.
Set beside the evidence view, the room can see where its views differ and why, and which of those differences carry the outcome.
Narrow it.
The discussion can move from people to the points that matter. Views may shift in either direction, and the decision that follows is one the committee owns.
The risk is measured. The room still disagrees.
When a serious risk reaches your committee, everyone at the table already holds a picture of it. The chief executive has one. So do operations, quality and the risk analyst. Those pictures rarely match, and the meeting argues about conclusions while the reasons behind them stay out of sight.
Most risk tools treat that as noise around the number. Perilometry treats it as part of the risk: how the people deciding see it, where their views part from the evidence, and whether their judgement was challenged and recorded before they acted.
After a major loss, investigations can find that the warning was already inside the organisation. It sat in one view that nobody put beside the others.
For example: the Rogers Commission on the Challenger loss (1986); the Columbia Accident Investigation Board (2003); and evidence given to the UK Treasury Committee on HBOS.
The problem is well known. It is rarely measured.
Fifty years of research point the same way: the people deciding on a risk see it differently, and unchecked judgement is where serious losses begin.
- 55%
- The median difference between two insurance underwriters pricing the same policy. Senior executives asked how much variation to expect said about 10%. Kahneman, Sibony and Sunstein, Noise, 2021.
- Fewer than 1 in 3
- The share of individual experts who were statistically accurate when tested on questions with known answers, across 33 studies. Colson and Cooke, 2018.
- 84 inquiries
- In a study of 84 British accident inquiries, disasters developed over long periods in which warning signals failed to be noticed. Turner, as reported in Choo, 2008.
Risk frameworks ask you to mind the human element. Perilometry makes it visible.
Established quantitative methods train the analyst to give honest ranges and combine them into one estimate. Perilometry does something else: it shows the decision-makers’ own views of the same quantified risk beside the evidence view. We know of no other approach that does.
What makes it different.
- It shows the gap, as well as the risk
- One risk is shown through several views side by side: senior roles, the business case, today’s process and an evidence view built from sources. The distance between how people see the risk and what the evidence supports is shown and explained. The aim is a better-founded view, never blame.
- People and evidence are kept apart
- Every view says what it is: taken from the people themselves, or reconstructed. Hindsight is labelled. No view is recorded as no view, never as zero. Disagreement stays visible and is not averaged away. The evidence view is an estimate too, and it is open to the same challenge as any other view.
- Judgement is recorded in its own right
- A committee freezes a baseline, tests assumptions on copies, and records two separate judgements: whether it agrees with the analysis, and what it will fund. In our worked examples, the appetite is set before any result is seen.
- No AI produces a number
- AI researches and drafts. It proposes inputs, each with its source and the exact words it rests on. Every figure is computed by a risk model, and documents quote only what the model produced. The ranges AI proposes can be challenged like any other input.
- An AI builder does the heavy work
- An AI builder will take a decision you describe and turn it into a governed case: it researches, drafts and challenges the work, and a person approves each step.
- Nobody marks their own work
- Each case, and each view within it, can be put before an AI-generated panel: a hostile reviewer who attacks it, a defender who answers and a judge who rules. Every fix is checked by a reviewer other than its author, and no value is changed after results are seen without a fresh, blind review. Review by independent human experts is the next step.
- A person holds the gates, on the record
- Nothing is compared until a person has reviewed the sources and approved. Who approved what, and when, can be shown.
- The system learns from its mistakes
- Every error found is traced to the rule that failed to prevent it, and the rule is changed, so that each case makes the next one harder to get wrong.
Your field’s language. One structure underneath.
Perilometry does not ask a quality head to talk like a security chief. Each field describes a risk in its own words, and the same rigorous structure sits beneath all of them.
A pharma quality team speaks of batch runs and contaminated batches. A bank’s financial-crime team speaks of programme deficiencies and enforcement. The words change; the structure does not. Worked examples exist in six fields: cyber, pharma, aviation safety, health technology, AI risk and financial crime. The risk factors in the structure underneath are mapped to the Open FAIR™ taxonomy.
For the people who answer for the decision.
- Boards and audit committees that need to know a risk was examined properly before it was accepted.
- Risk committees and chief risk officers who must bring several functions to one decision.
- Risk experts who brief senior people and need their analysis heard on its merits.
Perilometry is a quantitative risk platform with its own engine. It handles complex assessments: ranges in place of point estimates, rare catastrophic events, controls that cover only part of a threat, and losses that are insured or passed on. It runs that same analysis for every view in the room and for the evidence. It is built for depth on the decisions that matter most, one risk at a time, and is not a register for cataloguing every risk.
Built on ranges and evidence.
- Every figure shows its basis. You can see where each figure comes from.
- Ranges, not false precision. Uncertainty is shown, not hidden.
- Evidence you can challenge. A reviewer can question a source or record newer evidence. Nothing changes a result silently.
- Evidence informs judgment. The judgment stays with the people responsible for it.
A decision you can stand behind.
Before anything goes wrong
The record shows that the views in the room were surfaced, examined against the evidence and decided on deliberately.
If something does go wrong
It helps show which belief did not hold, so that recovery effort can go to the real weak point.
Bring one contested risk.
In a private demonstration we walk through a worked example, then discuss a risk your committee is facing.
Request a private demonstration or write to demo@perilometry.com