Why Use Scenario Prompts?
Routine reviews are valuable for finding familiar exposures: a process step that fails repeatedly, a control that is inconsistently applied, or a dependency that staff already recognize. They may be less effective for disruptions that are unusual, involve several conditions at once, or have not occurred within the organization’s recent experience. A structured what-if prompt deliberately asks people to consider those less familiar paths.
A scenario prompt is a focused question about how a specified condition could disrupt an objective. It is not a prediction that the condition will occur, a claim that it is likely, or a complete measurement of the resulting risk. Its purpose is to help participants identify a credible cause-event-effect pathway that ordinary process review may not reveal.
A useful prompt is specific enough to guide discussion, but open enough to reveal consequences participants did not anticipate. “What if something bad happened?” is too broad. “What if the organization’s primary claims platform were unavailable during a regional emergency, while access to its recovery environment was also delayed?” gives participants a starting condition and invites them to examine dependencies, duration, and effects.
The scenario should be challenging, not fanciful. Participants need not show that it has happened before, but they should be able to explain a credible mechanism by which it could happen. The aim is to identify a possibility worth examining, not to make a low-frequency event sound inevitable.
Build the Prompt Around a Clear Pathway
A practical prompt links four elements: the objective at risk, a trigger or initiating condition, the disruption that follows, and the possible effect on the objective. It can then be extended by asking what might make the disruption worse: a shared dependency, a delay in restoring service, limited alternative capacity, or simultaneous demand from several parts of the organization.
Identify what must continue or be achieved, such as issuing accurate claim payments by a required date.
Choose a specific condition, such as a regional power disruption, a supplier outage, or a previously unused system weakness.
Ask what stops working, who or what is affected, and how the disruption could impede the objective.
Test duration, dependencies, capacity, timing, and the availability of alternatives without assuming every failure occurs at once.
The prompt should make assumptions visible. If it assumes that a supplier is unavailable, that is different from assuming the supplier and the organization’s own backup both fail. If it assumes an outage lasts a day, participants should not quietly reason as if it lasts a month. Clear assumptions help the group explore a demanding scenario without confusing it with an unconstrained catastrophe.
A scenario can be developed by asking questions rather than adding more dramatic events. Is the organization dependent on one provider or location? Would the disruption occur during a period of unusually high workload? Could staff use a manual workaround, and for how long? Do customers, regulators, or counterparties depend on the same disrupted service? Each question tests a specific part of the pathway.
Worked Examples
Worked Example: A Claims Platform Outage During a Regional Emergency
Hypothetical problem. A property insurer relies on a hosted platform to receive claims and route them to adjusters. Its routine review finds no recent platform outage. Develop a scenario prompt that could reveal a low-frequency disruption and explain how the team should assess the historical evidence.
Method. Start with the service objective, then specify one credible disruption pathway. Treat any numerical probability as an illustration, not as an estimate derived from the scenario.
Step 1: State the objective. The relevant objective is to receive and route claims so adjusters can begin handling covered losses within the insurer’s intended service window.
Step 2: Form the prompt. “What if a regional emergency produces a sudden surge in claims while the hosted platform’s primary service region is unavailable? How would claims be received, prioritized, and routed if the recovery environment took longer than expected to become usable?”
Step 3: Trace the pathway. The trigger is the regional emergency. The event is a platform interruption during a claims surge. Possible effects include delayed claim intake, backlogs, difficulty prioritizing urgent cases, and slower service to policyholders. The scenario raises useful questions about whether alternative intake channels exist, whether staff can access customer information without the platform, and how much work those alternatives can handle.
Step 4: Interpret the absence of recent outages. Suppose, solely for illustration, that an annual event probability were 0.02, stayed constant, and each year were independent. The probability of observing no event in eight years would be:
The calculation shows that a quiet eight-year history can occur even under this illustrative event probability. It does not establish that the actual probability is 0.02; there is no basis in the problem data for estimating it. The history alone is therefore not evidence that the scenario is impossible.
Check. The prompt identifies an objective, a trigger, a disruption, and potential effects. It asks about recovery assumptions without asserting that an outage or a claims surge will occur together. The common mistake avoided is treating “no recent outage” as “no exposure.”
Worked Example: A Critical Data Supplier Becomes Unavailable
Hypothetical problem. An insurer uses an external supplier to obtain information needed for a renewal decision. Routine process reviews focus on errors in individual records. Develop a scenario prompt that tests a broader, infrequent disruption and identify the follow-up questions it should generate.
Method. Move from an individual transaction to a shared dependency. Specify the service affected, then examine the period and capacity within which alternatives could be used.
Step 1: State the objective. The objective is to complete renewal decisions using the information the insurer requires, within its operating timetable.
Step 2: Form the prompt. “What if the supplier’s service became unavailable during the peak renewal period following a disruption affecting its operating region? Which decisions would be delayed if the insurer could not obtain the required information, and what alternative sources or procedures could be used?”
Step 3: Trace possible effects. The event is an interruption to a shared information dependency. It could affect many renewal decisions at once, rather than cause errors in isolated records. Possible effects include delayed decisions, additional manual work, or decisions made with incomplete information if the process permits them.
Step 4: Ask bounded follow-up questions. Participants should establish which products and workflows use the supplier, whether the same supplier supports other critical activities, how long existing information remains usable, and how many cases an alternative process could handle. They should also distinguish between “an alternative exists” and “an alternative can handle peak-period demand.”
Check. The prompt reveals concentration and capacity questions without assuming that every product uses the same data or that all alternatives fail. The common mistake avoided is framing the issue only as a record-quality problem when the less frequent concern may be a broad service interruption.
Worked Example: A Delayed Recovery of an Actuarial Production Process
Hypothetical problem. A company’s actuarial team produces a required quarterly result using internal data feeds, a calculation environment, and a controlled release process. The team has not experienced a missed deadline. Write a scenario prompt that tests a low-frequency disruption and separate a useful scenario from an unsupported worst case.
Method. Anchor the prompt to the deadline and the production chain. Introduce a disruption that could affect a dependency, then ask about restoration and fallback capacity.
Step 1: State the objective. The objective is to produce and review the required result by its deadline using sufficiently reliable data and calculations.
Step 2: Form the prompt. “What if a failure in a shared data-feed service is discovered shortly before the quarterly production run, and replacement data cannot be validated immediately? How would the team determine whether it can recover the process and complete the required review by the deadline?”
Step 3: Identify the pathway. The trigger is a failure in a shared service. The disruption is delayed access to data that must be validated before production. Possible effects include a delayed run, compressed review time, or use of an alternative dataset that requires additional checks. The prompt does not assume that the calculation environment, data, and review controls all fail.
Step 4: Test plausibility and boundaries. Ask which outputs depend on the service, whether another validated source exists, how long validation usually takes, and what work can be performed while data access is restored. These questions reveal what would need verification. A claim that every source, system, and reviewer becomes unavailable indefinitely adds events without establishing a credible pathway, so it is less useful than the bounded prompt.
Check. The scenario is anchored to a real objective and a specific dependency while leaving the consequences open for investigation. The common mistake avoided is equating severity with usefulness: an extreme story is not automatically a better risk prompt.
What the Evidence Can and Cannot Establish
Recent experience can help establish which processes, dependencies, or controls deserve attention. It cannot, by itself, rule out a disruption that is infrequent, newly possible, or absent from the period observed. A basic probability result explains why. Assume an event has the same probability \(p\) in each period and that periods are independent. Then the probability of no events in \(n\) periods is:
This follows because the probability of no event in one period is \(1-p\), and independence means the probabilities of no event across periods multiply. The result does not supply \(p\), prove that periods are independent, or justify using it for a particular risk without evidence. Its limited lesson is that observing no event is compatible with a positive event probability.
Scenario prompts are therefore a way to test exposure and assumptions, not a substitute for estimating likelihood, assessing impact, or selecting a response. A scenario that merits follow-up may need evidence about dependencies, recovery arrangements, or operational capacity. If participants cannot identify a credible mechanism, that is a reason to refine or set aside the prompt—not to preserve an ever-expanding worst case.
Common Mistakes
- Using an unbounded prompt. “What if everything failed?” does not identify a pathway. Specify a trigger and ask what it could disrupt.
- Requiring a prior incident. A scenario can be worth exploring even when the organization has no direct experience of it.
- Turning a scenario into a forecast. A plausible pathway does not mean the event is expected or that its probability has been measured.
- Stacking assumptions without checking them. Multiple simultaneous failures may be considered, but each should have a reason to be included.
- Confusing a workaround with sufficient capacity. An alternative process may exist yet be unable to handle a prolonged outage or a surge in demand.
Check Your Understanding
For each problem, focus on making the scenario specific and bounded. The final answer gives one defensible response; other answers may also be reasonable if they state a credible pathway.
- A hypothetical insurer depends on one external service to deliver daily policy updates. No outage has occurred in the past three years. Write a prompt that tests a low-frequency interruption without treating the quiet history as proof of safety. (Answer: “What if the external service were unavailable during a high-volume update period? Which policy activities would be delayed, and what alternative process could handle the backlog?”)
- A claims team has a manual intake procedure for system outages. What additional question tests whether this is a meaningful contingency during a regional event? (Answer: Ask how many claims the manual procedure can process, and for how long, if intake volume rises while the system is unavailable.)
- A reporting team uses a data source that was introduced recently. It has no history of failure. Name one suitable prompt and one unsupported conclusion to avoid. (Answer: Prompt: “What if the new source were unavailable before a reporting deadline, and what validated alternative could be used?” Avoid concluding that the source cannot fail because no failure has occurred.)
- In a hypothetical independent-period model, an event has a constant probability of 0.10 per year. Calculate the probability of no event in three years, to four decimals, and explain what it means. (Answer: \(0.90^3=0.7290\); under the stated assumptions, no event in three years has probability 0.7290, so a quiet history does not imply zero event probability.)
- A proposed scenario says that a regional storm disrupts a critical supplier, the organization’s recovery site, all staff access, and every alternative vendor indefinitely. What is the main weakness, and how would you improve it? (Answer: It stacks several failures and an indefinite duration without establishing their pathways; specify one credible disruption and ask which dependencies, recovery constraints, or alternatives could extend its effects.)