CustomerNode™

Theme: Light
Financial Services › Insurance › Claims Operations

Catastrophe Response

Complex multi-party engagements where risk, regulation, and claim resolution require coordinated action.

CoreLogic RMS (Moody's) AIR Worldwide Verisk
Inside this journey

  1. Pre-Discovery

    Align the room on outcomes, decision process, and constraints before deeper discovery.

    1. Stakeholder Alignment

      Confirm decision roles (CRO, Head of Cat Risk, Chief Actuary, CIO, VP Claims), timelines, and success criteria for modeling and response decisions.

      Alignment Questions

      Quick Snapshot — Who’s in the Room?

      • Tell us your role and the primary team that will own this modeling and response relationship. Options: Chief Risk Officer, Head of Catastrophe Risk, Chief Actuary, CIO/Head of Technology, VP Claims/Head of Claims Ops, Portfolio/Underwriting Lead, Other
      • What is your target timeline for selecting and onboarding a catastrophe modeling + response partner? Options: Immediate (0–1 month), Near term (1–3 months), This quarter (3–6 months), 6–12 months, Undetermined/Exploratory
      • Who must sign off on modeling accuracy, commercial terms, and event activation decisions? Please list names and titles if possible.
      • What would success look like at contract-signing day from your perspective—what’s the single most important deliverable? Options: Validated model accuracy, Clear SLAs for live events, Integration plan with production systems, Pilot or POC commitment, Commercial terms only, Other
      • Is there anything about past vendor engagements that made you skeptical or enthusiastic? Share a short example.

      Show Me the Current Signal — Where Your Models Live Today

      • What if your current modeling setup is creating a false sense of precision—where might that be hiding?
      • Describe your current modeling workflows: which models you run, how often, and who owns each step.
      • Which data sources feed your exposure and claims estimates today? (select all that apply) Options: Policy administration (PAM), General ledger/finance, Claims system, Catastrophe modelling vendor feeds, Third-party data (satellite, weather), Geocoding vendor, Other
      • How automated are end-to-end runs from exposure ingestion to portfolio loss output? Options: Fully automated, Mostly automated with manual checks, Semi-manual, Fully manual/ad-hoc
      • When you run loss estimates today, how long does it take to produce a usable portfolio-level number? Options: <1 hour, 1–4 hours, 4–24 hours, 1–3 days, >3 days
      • What integrations exist between your catastrophe models and underwriting, pricing, or reinsurance systems? Please name systems and integration types.

      When the Storm Hits — What Breaks First?

      • When an event occurs, which single operational failure most undermines your ability to respond? Options: Data latency/quality, Insufficient adjusters, Slow model runtimes, Lack of governance/decision path, Inability to integrate real-time signals, Other
      • Describe the most recent significant event you managed: what surprised you operationally and what caused the biggest delay.
      • How do spikes in claim volume currently get triaged and routed—what are the manual handoffs or bottlenecks?
      • What percentage of claims in a major event do you expect to be handled by internal adjusters versus external partners? Options: 0–25%, 26–50%, 51–75%, 76–100%
      • Which failure modes (data gaps, model drift, staffing shortages, vendor SLAs) have recurring impacts on board/reinsurer confidence? Options: Data gaps, Model drift/unvalidated assumptions, Staffing shortages, Vendor/service reliability, Slow decision cycles, Other

      Who Decides and How Fast — Governance Under Pressure

      • If the board demanded a defensible 24-hour loss estimate, could your current governance deliver it—and if not, why? Options: Yes, confidently, Yes, with caveats, No, not without help, Unsure
      • What are the approval gates for publishing loss estimates internally and to external stakeholders (board, reinsurers)?
      • How are model ownership and change control structured—who approves model updates, inputs, and calibrations? Options: Actuarial/Model Governance, Risk Management, CIO/Data Team, External Vendor Control, Cross-functional committee, Other
      • How quickly can commercial or claims teams execute contract-level response actions (e.g., activate adjuster tiers, deploy drones) once an estimate is accepted? Options: Within hours, Same day, 1–3 days, >3 days
      • Who in your organization is empowered to make real-time activation decisions during an event?

      What Would True Confidence Feel Like?

      • Imagine standing before your board with numbers you fully trust—what specific tolerances, speeds, or assurances are you holding up as proof?
      • What accuracy tolerances (e.g., +/- %) are acceptable for: (a) early 24-hour estimates, (b) 7-day refined estimates, (c) final reserves? Options: <5%, 5–10%, 10–20%, >20%, Not defined
      • What runtime targets do you need for live-event model outputs to be useful operationally? Options: <15 minutes, <1 hour, 1–4 hours, 4–12 hours
      • Which KPIs will you use to judge success during and after an event (accuracy, time-to-first-estimate, claims cycle time, payout variance, customer satisfaction)? Select top three. Options: Estimate accuracy vs. eventual paid loss, Time-to-first-estimate, Claims handling SLA adherence, Adjuster deployment time, Customer satisfaction/NPS after event, Reinsurance recoverable accuracy, Other
      • How important is real-time situational awareness (satellite, drones) versus probabilistic model outputs for your decision-making? Options: Essential, Very important, Somewhat important, Nice to have, Not important

      Data & Integration — Is Your Exposure Healthy Enough?

      • Is your exposure data structured and geocoded to a level that supports high-confidence portfolio loss estimation—or are there critical gaps? Options: Fully structured & geocoded, Mostly structured with pockets of gaps, Significant gaps exist, We lack consistent exposure data
      • Which exposure attributes are consistently available and reliable for modeling (building value, occupancy, construction, deductible, policy limits, retro dates)? Options: Building value, Occupancy/use, Construction type, Policy limits/deductibles, Retro dates/coverage triggers, Other
      • How often are exposure feeds refreshed and validated (nightly, weekly, on-policy-change)? Options: Near real-time, Daily, Weekly, Monthly, Ad-hoc/manual
      • What systems will need direct integration for a production deployment (PAM, Claims, Reinsurance placements, BI/Reporting)? Please list.
      • Do you have security, legal, or PII constraints that limit how we might access or process policy-level data? Options: Standard contractual access permitted, Requires strict access controls/air-gapped processes, Data anonymization required, Legal or regulatory restrictions prevent some access, Unsure

      Operationalizing Response — From Estimate to Action

      • If you had a guaranteed T+6-hour operational play for a major event, what would the single most important activity be? Options: Publish a defended portfolio estimate, Activate primary adjuster tiers, Initiate satellite/drone imagery tasking, Notify reinsurers/placement teams, Open claims intake channels
      • How do you currently prioritize which claims or areas get adjuster or drone resources first? Options: Severity/probable loss, Concentration of exposure, High-value accounts, Regulatory/prioritization rules, Random/first-come
      • Describe your ideal SLAs for response partner actions (first estimate, dispatch time, imagery delivery, adjuster visit).
      • Which response capabilities are deal-breakers for you: licensed adjusters, drone ops, satellite analytics, mobile claims units, or local adjuster networks? Options: Licensed adjusters, Drone-based damage assessment, Satellite imagery analysis, Mobile claims processing units, Local/regional adjuster rosters, All of the above
      • Are there geographic limitations, carrier agreements, or regulatory constraints that would limit the use of certain response modalities (e.g., drones) in your footprint? Options: Yes — multiple restrictions, Some regions restricted, No significant constraints, Unsure

      Acceptance, Validation, and Trust — How Will You Say Yes?

      • What formal acceptance tests or validation steps must be satisfied before you will consider a model output trustworthy for financial decisions?
      • Which of the following would most increase your confidence in estimates: external benchmarking, back-testing on past events, parallel runs, or third-party audit? Options: Back-testing on past events, Parallel runs with existing models, External benchmarking vs peers, Independent third-party audit, All of the above
      • How do you want post-event validation to be delivered—summary KPIs, case-by-case reconciliations, or an executable playbook for continuous improvement? Options: Summary KPIs, Case-by-case reconciliations, Executable improvement playbook, Combination
      • What tolerance for model miss (e.g., under/over-estimation) would be operationally acceptable before you change provider or trigger remediation? Options: <5% average, 5–10%, 10–20%, >20% or dependent on event
      • How frequently would you want formal governance reviews of model performance—monthly, quarterly, after significant events, or another cadence? Options: Monthly, Quarterly, After every significant event, Annually, Other

      What’s Standing in the Way of Change?

      • What single organizational habit or procurement reality most prevents you from moving faster on model and response improvements? Options: Lengthy procurement cycles, Siloed ownership of data/systems, Budget cycle misalignment, Risk aversion to third-party models, Legal/data-sharing concerns, Other
      • Which stakeholders tend to resist change in this area and why—technical, commercial, cultural, or regulatory reasons? Options: Technical (integration/data), Commercial (cost/terms), Cultural (internal teams), Regulatory/compliance, Other
      • Have past pilots failed for reasons that could be addressed with a different structure (shorter scope, clearer acceptance criteria, shared risk)? Please describe briefly.
      • What internal approvals or evidence would make procurement and sign-off straightforward (POC results, executive sponsor, pilot ROI)? Options: POC results, Executive sponsorship, Clear ROI/Cost-benefit, Regulatory sign-off, Reinsurance alignment, Other
      • How comfortable are you with a staged approach (pilot → scale → production) to de-risk adoption? Options: Very comfortable, Somewhat comfortable, Prefer full deployment, Not comfortable

      Commitable Next Steps — What Would You Try First?

      • If we could agree to one small, low-risk pilot to prove value, what would that pilot have to deliver in 60–90 days?
      • Which pilot scope would you prefer: model back-test on historic events, live parallel run for a limited book, or a response readiness drill (drone/satellite + adjuster dispatch)? Options: Back-test on historic events, Live parallel run on limited book, Response readiness drill, Data integration proof-of-concept, Other
      • Who would be the internal point(s) of contact and what level of time commitment can they allocate to a pilot (hours/week)?
      • What commercial or legal barriers need to be resolved before a pilot can begin (data-sharing agreement, NDAs, SLAs, procurement approvals)? Options: Data-sharing agreement, NDA, Vendor SLAs, Procurement approval, Regulatory clearance, None
      • How would you like us to follow up—an executive summary, a technical plan, or a proposed SOW with timelines? Options: Executive summary, Technical plan, Proposed SOW/timeline, Schedule a working session, Other

    2. Current State Mapping

      Document existing modeling workflows, data sources, system integrations, claims response capacity, and failure modes that block goals.

      Current State

      Where We Are Today — A Quick Snapshot

      • To start light: in a single sentence, how would you describe your current catastrophe modeling and response setup?
      • Which of these best describes who owns modeling inputs and outputs today? Options: Risk/Analytics team, Actuarial, IT/CIO, Claims, Shared governance (cross-functional), Third-party vendor
      • What model types do you rely on today (select all that apply)? Options: Proprietary internal model, Commercial third‑party model, Open-source models, Hybrid ensemble, Rule-of-thumb / manual estimates
      • Roughly how fast do you need day‑one loss estimates during an active event (pick one)? Options: Within 1 hour, 1–6 hours, 6–24 hours, 24–72 hours, Longer / Not time‑sensitive
      • Who on your team typically fields executive or reinsurer questions during an event (roles)? Options: CRO, Head of Cat Risk, Chief Actuary, CIO/CTO, VP Claims, Head of Underwriting, Other

      What Keeps Your Numbers From Being Trusted?

      • When you see a meaningful gap between modeled loss and actual outcomes, what’s the first thing you suspect is wrong? Options: Exposure mapping errors, Policy terms/loading errors, Hazard representation, Vulnerability/fragility curves, Post-event reporting delays, Other
      • How often do you run back‑tests or event validation exercises against real losses? Options: Continuously / rolling, Quarterly, Annually, Irregular ad hoc, Never
      • Tell us about a recent validation that surprised you—what was wrong, and what did that feel like for leadership?
      • Which data issues most erode confidence in model outputs for your stakeholders? Options: Incomplete policy data, Incorrect geocoding, Out-of-date exposure attributes, Poor claims linkage, Inconsistent reinsurance layers, Other
      • When confidence drops, which decision gets delayed or changes most often (pricing, reserving, reinsurance purchase, claims triage)? Options: Pricing, Reserving, Reinsurance decisions, Claims resource allocation, Regulatory reporting, Other

      Who and What Are Tied into Your Models?

      • If your model were a central nervous system, which external systems are its nerves—what systems feed it in real time? Options: Policy administration (PAM), Claims management system (CMS), GIS / mapping, Telemetry / IoT, Reinsurance placement systems, Data warehouse / MDM, None of the above
      • Provocative pause: what would break first if one of those integrations failed during a storm? Options: Loss run refresh, Policy lookup & terms, Geolocation matching, Claims intake / duplication detection, Reinsurance reporting, Other
      • Which integration(s) are currently automated vs. manual? Options: Fully automated API feeds, Scheduled batch extracts, Manual CSV uploads, Ad hoc for events only, No integrations
      • How frequently are exposure and policy attribute feeds refreshed outside events? Options: Real‑time/near‑real‑time, Daily, Weekly, Monthly, Less often
      • Please list any third‑party data providers or internal master data sources that are mission‑critical to model accuracy.

      When Events Happen, Where Do Bottlenecks Live?

      • During an active catastrophe, what single process most commonly slows your ability to produce actionable loss estimates? Options: Data ingestion and cleaning, Exposure geocoding, Model run times/compute limits, Cross-team signoff/governance, Claims field validation, Reporting and distribution
      • Which of these capacity limits have forced you to change your real‑time approach in past events? Options: Insufficient compute for ensemble runs, Not enough licensed adjusters, Data throughput/API throttling, Lack of on‑call analytics staff, Vendor SLA constraints, Other
      • Describe a recent event where you missed your internal timeline—what happened and who had to compensate?
      • How do you currently prioritize claims triage geographically and by exposure during surge conditions? Options: Severity-first, Density/accumulation-first, VIP/line-of-business focused, Random/first-come, Other
      • Who signs off on day‑one estimates for external stakeholders (board, reinsurers)? Options: CRO, Chief Actuary, Head of Cat Risk, VP Claims, CIO, Cross-functional committee

      What Breaks When You Need Answers Fast?

      • If you had to name the Achilles’ heel that surfaces under time pressure, what is it? Options: Inconsistent policy data, Slow model compute, Lack of field validation, Conflicting cross‑team numbers, Poor scenario governance, Other
      • How often do manual workarounds (spreadsheets, adhoc scripts) replace your production process during a crisis? Options: Always, Often, Sometimes, Rarely, Never
      • When a workaround is used, who bears the operational risk and how long does it typically take to revert to standard process?
      • Which single automation, if implemented today, would reduce your time‑to‑estimate the most? Options: Real‑time exposure feed, Automated geocoding, Scalable cloud compute for ensemble runs, Integrated claims intake, Pre-built dashboards for stakeholders
      • Tell us about a near‑miss or failure during an event and what you learned from it.

      How Do You Measure Readiness (and Why It Feels Risky)?

      • When you say you are ‘ready’ for a catastrophe, what tangible indicators must be true? Options: Fresh exposure data loaded, Model ensemble validated, Adjuster roster on call, Reporting dashboards functional, Stakeholder governance aligned
      • What KPIs or SLAs do you track that relate specifically to modeling and response readiness? Options: Time-to-first-estimate, Model run completion rate, Geocoding accuracy, Adjuster dispatch time, Claims handling SLA, Other
      • Where do your current KPIs fail to capture the true operational risk?
      • How does it feel internally—are teams defensive, collaborative, or resigned—when readiness metrics miss targets? Options: Defensive/conflict, Collaborative/problem-solving, Resigned/accepting, Surprised/uncertain
      • Which governance or escalation paths do you wish were faster or clearer during model disagreements?

      If You Could Wave a Wand, What Would Be Different?

      • Imagine day‑one of the next event: what single change would make you feel calm rather than rushed?
      • Which capabilities would you most want from a partner to close current gaps (select up to three)? Options: Real‑time exposure integration, Pre‑validated model ensemble, On‑demand compute scaling, Adjuster & drone dispatch, Satellite analytics, Governance and runbooks
      • How would your organization measure success after that change—what outcomes would be different?
      • Which internal stakeholders would need to be convinced first, and what would their main objection be? Options: CRO, Chief Actuary, CIO, VP Claims, Finance, Board/Exec
      • If you had one small pilot you could run with a vendor in the next 60 days, what would you pilot? Options: End‑to‑end day‑one estimates, Automated geocoding refresh, Adjuster surge program, Satellite/drones for a region, Model validation against a known event

      Next Steps — What Would Make This Mapping Actionable?

      • What specific artifacts would you expect after this discovery to feel we’ve done useful work (select all that apply)? Options: Current-state data flow diagram, Mapping of failure modes & mitigations, Priority backlog for integrations, Pilot plan with milestones, RACI for event activation
      • Who should be involved in a 90‑minute follow-up workshop to validate our map (name roles)? Options: CRO, Head of Cat Risk, Chief Actuary, CIO/IT integration lead, VP Claims, Data Ops/MDM, Other
      • Which quick wins could we realistically deliver in 30–60 days to reduce your biggest pain? Options: Automate one exposure feed, Run a model validation report, Provision adjuster surge list, Deploy a day‑one dashboard prototype, Other
      • What constraints (procurement, security, data sharing) would block immediate progress we need to know about now? Options: Legal/data sharing limitations, Security/compliance reviews, Procurement cycles, Budget cycle timing, Executive approval required, None
      • Finally, what would make you feel confident that this discovery accurately reflects reality rather than a best‑case picture?

  2. Outcome Discovery

    Define target outcomes—model accuracy tolerances, live-event runtime targets, staffing SLAs, and measurable success signals.

    Discovery Questions

    Quick Grounding — Who’s in the Room?

    • Which role(s) are you representing in this conversation? Options: Chief Risk Officer, Head of Catastrophe Risk, Chief Actuary, CIO/Head of Data, VP Claims / Head of Claims Operations, Modeling/Analytics Lead, Other (please specify)
    • Tell us briefly about the scale of the portfolio you’re focused on for catastrophe planning (policies, GWP, or insured value—whichever you track).
    • Which perils and lines of business matter most for your catastrophe exposure today? Options: Hurricane (wind & flood), Earthquake, Wildfire, Flood (river/flash), Severe convective storms/hail, Commercial property, Personal lines (homeowners), Other
    • How do you primarily use catastrophe model outputs in your day-to-day decisions? Options: Pricing/underwriting, Portfolio accumulation management, Reinsurance buying/structure, Regulatory capital and reporting, Claims surge planning, Reserve setting, Other
    • Who ultimately signs off on live-event loss estimates and external communications at your firm? Options: CRO, Head of Cat Risk, Chief Actuary, CIO, VP Claims, CEO/Board, Other

    If the Board Asked for a Number Right Now, What Would You Say?

    • When a major event occurs, how confident are you that your initial loss estimate will hold up after 30 days? Options: Extremely confident (±5% or better), Confident (±10%), Somewhat confident (±20%), Low confidence (more than ±20%)
    • How long does it typically take your team to produce a first credible portfolio loss estimate after an event begins? Options: Under 1 hour, 1–6 hours, 6–24 hours, 24–72 hours, More than 72 hours
    • What internal pressures influence that timeline (board, reinsurers, regulators, distribution partners)? Please rank the top two. Options: Board/Executive leadership, Reinsurers/ retrocessionaires, Regulators, Investor/Parent company, Wholesale/broker partners, Policyholder communications
    • Describe a recent live event where timing or confidence of estimates caused material pain—what happened and who was most affected?
    • Which stakeholder outcomes are most time-sensitive for you during an event (choose up to three)? Options: Time-to-first-estimate, Estimate accuracy vs. final loss, Geographic granularity of loss, Confidence intervals / uncertainty quantification, Speed of adjuster deployment, Reinsurer communication

    What’s the Worst Surprise Your Models Ever Delivered?

    • Tell us about a time your model outputs materially surprised leadership—what was the surprise and what did it cost?
    • Which types of model failure have hurt you most: systematic bias, underestimation in tails, spatial misallocation, or timing/latency issues? Options: Systematic bias (consistent over/under), Tail underestimation (rare events), Spatial misallocation (wrong locations), Runtime/latency problems, Data integration errors, Other
    • How do you detect those failures today—audits, post-event reconciliation, reinsurer stress tests, regulators, or anecdotal feedback? Options: Automated back-testing, Manual post-event reconciliation, Reinsurer feedback, Regulatory review, Claims/adjuster feedback, We rarely detect them early
    • When a model error is discovered, what downstream actions are triggered and how long do they usually take (reserve adjustments, communication, reinsurance claims)?
    • How much of the surprise is technical (model/math) versus operational (data, people, approvals)? Please estimate percentages. Options: Mostly technical (>75%), More technical than operational (51–75%), Even split (~50/50), More operational than technical (51–75%), Mostly operational (>75%)

    What Quiet Frictions Slow Your Response?

    • Which internal friction do you believe causes the biggest delay during activation—data access, approvals, staffing, vendor coordination, or communications? Options: Data access/quality, Internal approvals/governance, Staffing/adjuster availability, Vendor coordination (imagers/drones), External communications/PR
    • How frequently do data permissions or legal concerns block real-time model runs or external data sharing in a crisis? Options: Almost always, Often, Sometimes, Rarely, Never
    • Give a specific example where an operational handoff failed (e.g., model handed to claims but claims couldn’t act). What broke down?
    • Which part of your tech stack slows down live-event modeling the most? Options: Policy admin system, Claims system, Data lake/warehouse, APIs and integrations, Manual spreadsheets, Other
    • If you could remove one operational friction instantly, which would it be and why?

    If You Could Name One Acceptance Test for Models, What Would It Be?

    • Which accuracy metric matters most to you when judging model acceptability? Options: Mean absolute error (MAE), Mean absolute percentage error (MAPE), Bias at the portfolio tail (P99/P95), Hit-rate of high-loss locations, Concordance with claims within 30 days, Other
    • What numeric tolerance would you accept for first-day estimates for typical events (pick best match)? Options: Within ±5%, Within ±10%, Within ±20%, Within ±30%, No firm tolerance—depends on event
    • How often would you require updates to those estimates during the first 72 hours of an event? Options: Continuous/real-time, Every hour, Every 3–6 hours, Twice daily, Daily
    • Which service-level agreements for staffing and response would you expect to be contractually guaranteed? Options: Time-to-adjuster-dispatch, Drone imagery turnaround, Satellite analysis delivery, Analyst availability for query, API uptime/latency
    • Please list any non-negotiable acceptance criteria we should be aware of (regulatory thresholds, reinsurance triggers, board rules).

    Describe the Two Numbers You’d Want on Day One

    • If you could get only two pieces of information within 24 hours post-event, what would they be and why?
    • What level of geographic granularity do those numbers need (portfolio, region, county, postal code, lat/long clusters)? Options: Portfolio-level only, Region/state, County/municipality, Postal code/zip, Property-level/lat-long
    • Do you need accompanying uncertainty information (confidence intervals, scenario bands) with those numbers? Options: Yes—required, Nice to have, Not necessary
    • How would you like those numbers delivered (dashboard, PDF report, API push, direct report to reinsurers)? Options: Interactive dashboard, PDF executive summary, API push to internal systems, Automated email to stakeholders, Direct report to reinsurers/partners
    • Who should receive those initial outputs within your organization (names/titles) and who must be copied externally?

    Tradeoffs — Speed, Accuracy, and Cost

    • If forced to choose during an activation, which would you prioritize? Options: Speed of first estimate over accuracy, Higher accuracy over speed, Balanced approach (minimum standards for both), Depends on event type
    • What level of accuracy degradation would be acceptable to gain a 4x improvement in speed (select one)? Options: < 5% degradation, 5–10%, 10–20%, >20%—not acceptable
    • Are you open to staged releases (fast initial estimate, then refined estimates) as a formal process? Options: Yes—this is ideal, Maybe—would need governance, No—one authoritative number only
    • How much additional budget or commercial flexibility would you consider to shorten time-to-first-estimate materially? Options: Willing to pay a premium, Prefer reallocation within current budget, Only if cost-neutral, Not willing
    • Which cost-of-error is more painful for you: underestimating losses or overestimating losses? Please explain.

    People, Playbooks, and Who Gets Activated

    • Who are the must-have roles on your incident command during a live catastrophe (choose up to four)? Options: Incident Commander (executive), Modeling lead/analyst, Claims leader, Reinsurance/retro lead, Communications/PR, Legal/compliance, IT/integration lead
    • What is your expected SLA for adjuster dispatch after an event for high-priority regions? Options: Within 4 hours, 4–12 hours, 12–24 hours, 24–48 hours, Depends on access/logistics
    • Describe your current runbook for activation—who authorizes it, and how often is it exercised? Options: We have a documented runbook, exercised annually, Documented, exercised sporadically, Informal runbook only, No runbook
    • Which response services matter most to you in the first 72 hours: adjusters, drones, satellite imagery, mobile claims units, or field leadership? Options: Licensed adjusters, Drone-based aerial assessment, Satellite imagery analysis, Mobile claims processing units, Field leadership/triage teams
    • How do your teams prefer to coordinate during activations (Slack/MS Teams, dedicated command center, email, phone, vendor portal)? Options: Slack/MS Teams, Dedicated incident command center, Email, Phone bridge, Vendor portal/portal integrations

    The Data Dependencies That Break or Make You

    • Which single external data feed, if lost, would most undermine your confidence in estimates? Options: Policy-level exposure data, Claims intake feed, Real-time hazard sensor data (wind gauges, seismic), High-res exposure geocoding, Satellite imagery
    • How often are your exposure and policy files refreshed for modeling purposes? Options: Real-time/near-real-time, Daily, Weekly, Monthly, Less often
    • What integration methods do you prefer for live events (API push/pull, SFTP batch, secure file share, manual upload)? Options: API (preferred), SFTP/batch, Secure portal/manual upload, Direct DB replication, Other
    • Do you have restrictions on sharing policy-level data with vendors during an event (anonymization, aggregate-only, legal approvals)? Options: Full policy-level sharing allowed, Policy-level with NDA and controls, Aggregate-only sharing, Requires legal approval per event, We do not share data externally
    • Tell us about any historical data quality issues (mismatched geocodes, missing limits, legacy policy mappings) that have affected model runs.

    What a Successful Pilot or Trial Looks Like

    • What primary metric would make a pilot with us a clear win for you (accuracy, speed, operational fit, TCO)? Options: Accuracy vs. final loss, Time-to-first-estimate, Operational integration ease, Adjuster/response speed, Total cost of ownership (TCO)
    • What duration and scope would you expect for a meaningful pilot (number of events simulated, portfolio slice, live-event trial)? Options: Short proof-of-concept (2–4 weeks), Multi-scenario pilot (1–3 months), Live-event readiness drill (scheduled drill), Season-long partnership (6–12 months)
    • List the top three acceptance criteria for a pilot to be considered successful (e.g., <10% MAPE day-1, <6-hour API latency, adjuster dispatch within 12 hours).
    • Who needs to approve a pilot and who signs the final commercial agreement? Options: CRO, Head of Cat Risk, Chief Actuary, CIO, Head of Claims, Procurement/Legal
    • What are the main internal blockers that could prevent you from executing a pilot in the next 90 days? Options: Budget/Procurement cycles, Legal/data sharing constraints, Resource availability, Regulatory approval, Lack of executive buy-in

    Decisions, Timing, and Who Moves the Needle

    • What does your decision-making timeline look like for a new modeling & response contract—from pilot to signature? Options: Under 1 month, 1–3 months, 3–6 months, 6–12 months, Longer than 12 months
    • Which stakeholders must be convinced for a go/no-go decision, and what are their primary concerns? Options: CRO (risk posture), CIO (integration), Chief Actuary (accuracy/reserves), VP Claims (operational fit), Procurement/Legal, Finance
    • If we could deliver one guarantee to accelerate your decision, what would be most persuasive (SLA, pilot performance, indemnity, integration timeline)? Options: Guaranteed accuracy SLA, Guaranteed time-to-first-estimate SLA, Contractual indemnity, Fixed integration timeline, Flexible commercial terms
    • Realistically, what would make you say 'yes' within your next budget cycle?
    • Is there anyone else we should include in this discovery to make sure the acceptance criteria reflect reality (names/titles)?

  3. Solution Experience

    Translate the customer’s portfolio and event scenarios into a shared view of how our modeling and response capabilities will deliver the defined outcomes.

    Experience Meetings

    • Pre-Work & Current State Confirmation
    • Scenario Modeling Workshop (Diagnosis -> Proof)
    • Response Capabilities Mapping Workshop (Proof -> Validation)
    • Joint Validation, Acceptance Criteria & Next Steps
    • Customer legal and procurement teams to prepare a list of required items for Mutual Commit (data sharing, SLAs, governance).
    • Seller to deliver scenario run reports (policy-level and aggregate) and visualizations within 48 hours.
    • Customer to review and confirm any policy-mapping exceptions or corrections identified during the run.
    • Both teams to log identified model or data gaps into a shared tracker with owners and target close dates.
    • If integration constraints surfaced, schedule a Technical Deep Dive to resolve API/data mapping issues.
    • Recap Modeled Outputs and Key Decision Thresholds
    • Agree on concrete response tiers, resource requirements, and SLA commitments tied to modeled thresholds.
    • Validate integration points for drone and satellite data into claims triage to reduce on-site inspections.
    • Document runbook actions and owners for each trigger level for later inclusion in Solution Scope.
    • Surface and quantify any incremental costs or logistical constraints for tiered response execution.
    • Seller to draft response runbooks for Tier 1–3 activations mapped to model thresholds and deliver to customer.
    • Customer to confirm adjuster roster availability, credentialing needs, and preferred vendors for drone/satellite services.
    • Both teams to align on SLAs (e.g., initial estimate time, dispatch SLA) and record any deviations from standard practice.
    • Estimate incremental operational costs for each response tier and circulate for finance review.
    • Review of Evidence: Modeling + Response Outputs
    • Agree and document a clear acceptance matrix (KPIs, SLAs, thresholds) that will be used to validate the solution in production and during live events.
    • Designate sign-off authorities and an escalation path for activation and dispute resolution during an event.
    • Schedule and scope a tabletop/drill to validate end-to-end behavior prior to Deployment readiness.
    • Confirm next steps and owners to move outputs into Solution Scope and Mutual Commit stages.
    • Create and share the acceptance matrix (KPIs, thresholds, measurement cadence) for stakeholder review and signature.
    • Schedule the tabletop/drill with required attendees and define the drill script and success criteria.
    • Seller to produce a Solution Scope draft (modeling modules, integrations, runbooks, SLAs) based on validated outputs.
    • Introductions & Objectives
    • Produce a single-sentence current state agreed by all attendees.
    • Quantify the operational and financial consequences of the current state for the selected portfolio slice.
    • Lock the exact portfolio extracts, event scenarios, and required SMEs/access for subsequent sessions.
    • Ensure pre-work deliverables and timelines are accepted by both teams.
    • Customer to deliver sample portfolio extract (policy-level) for selected slices with data dictionary.
    • Seller to provide data ingestion checklist and anonymization guidance.
    • Assign SME points-of-contact (Modeling, Claims Ops, IT) and confirm availability for workshops.
    • Agree and circulate the 2–3 event scenarios with scenario descriptions and assumptions.
    • Recap Current State, Consequence, and Future State
    • Prove that model outputs meet or identify where they miss the customer's defined accuracy tolerances and runtime targets.
    • Validate that modeled loss curves, claim counts, and uncertainty bands map to the customer's success signals.
    • Identify data or configuration gaps and assign remediation actions with deadlines.
    • Secure customer confirmation on the model-run methodology and acceptance criteria to be used later in acceptance checks.
    • Claims Volume to Resource Mapping
    • Define Acceptance Criteria & KPI Matrix
    • One-sentence Current State Readback
    • Scenario Inputs & Assumptions
    • Technology-Assisted Triage (Drones & Satellite)
    • Consequence Quantification
    • Decision Gates & Escalation Paths
    • Model Execution & Runtime Demonstration
    • Output Walkthrough: Loss Estimates & Uncertainty
    • Portfolio & Data Inventory Review
    • Simulated Drill / Tabletop Plan
    • Runbook Walkthrough & SLA Definitions
    • Cost, Logistics & Escalation Implications
    • Transition Plan to Solution Scope & Mutual Commit
    • Event Scenarios Selection
    • Sensitivity & What-if Adjustments
    • Final Q&A and Sign-off Intents
    • Validation Check: Is This Operationally Feasible?
    • Validation & Forced Confirmation
    • Pre-work & Access Checklist
    • Next Steps & Logistics
    • Gap Identification & Immediate Next Steps

  4. Solution Scope

    Define modeling modules, data integrations, reporting, response services (adjusters, drones, satellite), SLAs, runbooks, and acceptance criteria.

    Scope Configuration

    • Run Probabilistic Hurricane Simulation
    • Run Probabilistic Earthquake Simulation
    • Run Probabilistic Wildfire Simulation
    • Generate Per-Policy Loss Estimates
    • Produce Portfolio Accumulation Heatmap
    • Generate Real-Time Event Loss Footprint
    • Deliver Reinsurance Attachment Impact Report
    • Provide Claims Volume and Triage Projections
    • Deploy Licensed Adjusters to Impact Zones
    • Perform Drone Aerial Damage Surveys
    • Deliver Satellite Imagery Change Detection
    • Deploy Mobile Claims Processing Unit Onsite
    • Integrate Loss Outputs via API to Carrier Systems

    Scope Questions

    Run Probabilistic Hurricane Simulation

    • Which geographic regions should hurricane simulations cover? Options: US Gulf Coast, US East Coast, Caribbean, Central America, Mexico, Other
    • What return periods or event severities are required for analysis? Options: 10-year, 50-year, 100-year, 250-year, Custom
    • What exposure aggregation level do you need for hurricane outputs? Options: Per-policy, Per-building, ZIP/Postal, Census tract, County/State, Custom
    • Do you require temporal wind-field time-series (track-based) or only peak-loss snapshots? Options: Time-series (temporal), Peak-loss snapshots only, Both
    • Are there specific vulnerability or damage functions we should apply (carrier-provided or standard curves)? Options: Use carrier-provided curves, Use vendor-standard curves, Apply mixed/custom rules, No preference
    • List any regulatory, rating-agency, or internal assumptions that must be enforced (e.g., exposure exclusions, policy aggregation rules).

    Run Probabilistic Earthquake Simulation

    • Which tectonic regions or countries should earthquake modeling cover? Options: California/West US, Central/Eastern US, Japan, Europe, Global, Other
    • Which return periods and intensity measures are required (e.g., PGA, PGV, spectral acceleration)? Options: 10-year, 50-year, 100-year, 250-year, PGA, PGV, Spectral Acceleration, Custom
    • Do you require site-specific site-response or soil amplification adjustments? Options: Yes, provide site amplification, No, use regional GMPE only, Apply where available
    • What vulnerability/fragility datasets should be used for building classes and coverages? Options: Carrier vulnerability curves, Industry-standard fragilities, Hybrid/custom, Not applicable
    • What output granularity and formats are required (e.g., per-policy CSV, GeoTIFF, GIS layers)? Options: Per-policy CSV, Aggregate summaries, Shapefiles/GeoJSON, Raster outputs (GeoTIFF), Other
    • Are there special scenario sets (e.g., deterministic scenario list, aftershock sequences) to include? Options: Deterministic scenarios, Stochastic catalog only, Aftershock sequences, Custom scenario list

    Run Probabilistic Wildfire Simulation

    • Which regions and seasons should wildfire simulations cover? Options: Western US (California), Southeast US, Australia, Mediterranean, Global, Other
    • Do you require inclusion of dynamic fuel/vegetation maps and recent burn history? Options: Yes, include fuel maps and burn history, No, use static fuels, Include if available
    • What outputs are required: perimeter probability, flame length, burn probability, or per-policy loss estimates? Options: Perimeter probability, Burn probability, Flame length/behavior, Per-policy losses, All of the above
    • Do you need modeling of ember spotting or ember-driven spread across barriers (urban interfaces)? Options: Yes, ember spotting required, No, Only for specified zones
    • What temporal resolution and simulation horizons are needed (e.g., daily, hourly, seasonal)? Options: Hourly, Daily, Seasonal, Event-based
    • Are there suppression/mitigation policy assumptions to include (e.g., firebreaks, resource response)? Options: Model suppression effects, Ignore suppression, Apply simple reduction factor

    Generate Per-Policy Loss Estimates

    • Which policy fields must be mapped for per-policy outputs (e.g., limit, deductible, building value, construction class)?
    • Which coverages should be modeled separately (e.g., building, contents, BI/ALOP, roofing sublimits)? Options: Building, Contents, Business Interruption, Additional Living Expenses, Sublimits (roof, glass), Other
    • What rules should govern limit/deductible application (per-location aggregation, policy aggregate, per-occurrence)? Options: Per-location, Policy-aggregate, Per-occurrence, Custom rules
    • What output format(s) do you require for per-policy losses (CSV, JSON, direct API push)? Options: CSV, JSON, Direct API push, Database/SQL export
    • Do you require rounding, currency conversion, or reserve estimate fields in outputs? Options: Include rounding rules, Include currency conversion, Include reserve estimate fields, None
    • What acceptance criteria must per-policy outputs meet (e.g., reconciliation tolerance to expected samples)?

    Produce Portfolio Accumulation Heatmap

    • At what spatial aggregation level should accumulation be visualized (e.g., grid cell, ZIP, county, custom polygons)? Options: Grid cell (km), ZIP/Postal, County/State, Custom polygons (e.g., ceded zones)
    • Which metric(s) should the heatmap display (exposed value, sum insured, expected loss, AAL, peak loss)? Options: Exposed value, Sum insured, Expected loss, Average annual loss (AAL), Peak loss
    • Do you need multiple attachment point overlays for facultative or treaty analysis? Options: Yes, multiple attachment overlays, No, Single attachment only
    • What clipping or threshold rules should be applied for visualization (e.g., hide cells < $X)? Options: Hide below threshold, Show all cells, Highlight top X%
    • Which export formats are required for heatmap deliverables (interactive web map, PNG, GeoTIFF, CSV)? Options: Interactive web map, PNG, GeoTIFF, CSV, Shapefile/GeoJSON
    • Are there stakeholder-specific views or permissions needed (e.g., underwriter vs. catastrophe team)? Options: Yes, role-based views, No

    Generate Real-Time Event Loss Footprint

    • What is the required data refresh cadence during an active event (e.g., every 5 min, hourly)? Options: Every 5 minutes, Every 15 minutes, Hourly, Custom
    • What maximum end-to-end latency is acceptable from data receipt to reported footprint? Options: <5 minutes, <15 minutes, <1 hour, Up to 4 hours
    • Which triggering sources will start the real-time workflow (e.g., NHC advisory, seismic event, carrier notification)? Options: Official advisories (NHC/USGS), Carrier-triggered, Third-party alerts, Automated sensor feeds
    • Which output products are required in real-time (per-policy loss push, geospatial footprint, claims triage list)? Options: Per-policy push, Geospatial footprint, Claims triage list, Executive summary
    • What acceptance criteria and QA checks must run before each real-time push (e.g., data completeness, reconciliation thresholds)?
    • Are automated notifications or dashboards required for internal stakeholders/reinsurers during the event? Options: Yes: email/SMS/Webhook, No, Dashboard only

    Deliver Reinsurance Attachment Impact Report

    • Which treaty types and layers should be included (working cover, XS, aggregate stop-loss, facultative)? Options: Working cover, Excess of loss (XS), Aggregate stop-loss, Facultative, Other
    • What attachment points, limits, and reinstatement terms must be modeled?
    • Should the report include probabilistic exceedance curves, PML, and secondary per-event distributions? Options: Include all metrics (PML/AEP/OEP), Only PML and AEP, Custom set
    • What currency and aggregation rules should be used for treaty accounting and reporting? Options: USD, Local currency, Multi-currency with conversion, Other
    • Who are the intended recipients and what format do they prefer (underwriter PDF, cedant CSV, reinsurance broker portal)? Options: PDF executive report, CSV/data extracts, Broker portal package, API transfer
    • Are stress/test scenarios required (e.g., correlated events, market stress) and if so which ones? Options: Yes, specific stress scenarios, No, Standard set only

    Provide Claims Volume and Triage Projections

    • Which lines of business should projections cover (e.g., personal homeowners, commercial property, auto)? Options: Personal homeowners, Commercial property, Auto, Flood, Other
    • What horizon and cadence for projections do you need (first 24-72 hours, first 30 days, ongoing weekly)? Options: 0-72 hours (immediate), 0-30 days, 0-90 days, Ongoing weekly
    • Do you require triage categories and routing rules (e.g., field adjuster, remote assessment, immediate referral)? Options: Yes, define triage categories, No, provide counts only, Basic triage only
    • What granularity is required for staffing projections (by county, ZIP, adjuster-day demand)? Options: County, ZIP/Postal, Adjuster-day, State
    • Are expected accuracy tolerances or confidence intervals required for projections? Options: Yes, 90% CI, Yes, 95% CI, No tolerance required
    • Which output formats are needed for triage lists and routing (API push to claims system, CSV, dashboard)? Options: API push to claims system, CSV export, Dashboard visualizations, Email reports

    Deploy Licensed Adjusters to Impact Zones

    • What triggers mobilization of adjuster teams (e.g., threshold losses, carrier request, automatic dispatch)? Options: Automatic dispatch on threshold, Carrier-requested mobilization, Mutual aid/partner activation
    • What is the target number of adjusters or adjuster-days to pre-stage or deploy? Options: Small (1-10), Medium (11-50), Large (50+), Custom number
    • Are there state licensing, certification, or language requirements for adjusters? Options: State-specific licenses required, Language/translation required, No special requirements
    • Do you require adjusters to follow carrier-specific triage/runbook procedures or use our standard process? Options: Carrier-specific runbooks, Use vendor standard process, Hybrid
    • What duration and rotation cadence should be planned for deployed teams? Options: 1-3 days, 4-7 days, 2+ weeks, Custom rotation
    • Are logistics support items required (staging locations, travel, housing, security)? Options: Yes, vendor arranges logistics, Carrier provides logistics, Hybrid

    Perform Drone Aerial Damage Surveys

    • Which areas and priority zones should drone surveys cover? Options: Urban impacted areas, Rural/remote zones, Critical infrastructure, Carrier-selected hotspots, Other
    • What spatial resolution and deliverables are required (images, orthomosaics, 3D models)? Options: High-res images, Orthomosaic/stitched imagery, 3D point clouds/mesh, Video
    • Are there airspace or permitting restrictions we should plan for (FAA waivers, no-fly zones)? Options: Yes, expect restrictions, No known restrictions, Carrier will provide permissions
    • What turnaround time is required from flight to analyzed deliverables? Options: <24 hours, <48 hours, <72 hours, Custom
    • Do you require integration of drone outputs into carrier workflows (per-policy tagging, claims attachments)? Options: Yes, integrate via API, Provide files only, Dashboard access only
    • Any privacy, PII, or property owner consent requirements to observe? Options: Yes, consent required, No, Carrier will handle consent

  5. Mutual Commit

    Finalize commercial terms, data-sharing agreements, SLAs, governance, and escalation paths for live-event activation and ongoing use.

    Agreement Modules

    • Statement of Work (SOW)
    • Master Services Agreement (MSA)
    • Service Level Agreement (SLA)
    • Data Sharing & Use Agreement
    • Data Processing Agreement (DPA)
    • Security & Privacy Addendum
    • Commercial Schedule & Pricing
    • Payment Terms & Invoicing
    • Insurance, Indemnity & Liability Schedule
    • Integration & API Access Agreement
    • Live Event Activation & Escalation Plan
    • Adjuster & Response Services Addendum
    • Acceptance Criteria & Validation Checklist
    • Governance & Steering Committee Charter
    • Change Order & Amendment Process
    • Renewal, Termination & Exit Assistance
    • Regulatory Compliance & Audit Rights

  6. Deployment

    Operationalize rollout with readiness checks, enablement, and outcome validation.

    1. Pre-Deployment Readiness

      Confirm data feeds, test environments, API access, adjuster rosters, and incident command roles are provisioned and tested.

      Readiness Questions

      What's on Your Plate Right Now?

      • Which role are you representing in this conversation? Options: Chief Risk Officer (CRO), Head of Catastrophe Risk, Chief Actuary, CIO / Head of Tech, VP Claims / Head of Claims Operations, Reinsurance Lead, Other (please specify)
      • Briefly describe the single most important objective you need catastrophe modeling and response to deliver this year.
      • Which business decisions rely most heavily on your catastrophe model outputs right now? Options: Pricing / rate setting, Portfolio accumulation management, Reinsurance purchasing, Regulatory capital / reserving, Claims triage & staffing, Catastrophe planning & contingency, Other
      • How do you and your leadership currently judge whether a model or response partner is performing well? Options: Estimate accuracy vs actuals, Speed/latency of estimates, Adjuster fill-rate / response speed, Integration reliability, Clarity of assumptions / explainability, Cost / ROI, Other
      • Tell us about a recent event where model output materially affected pricing, reserving, or response—what worked, and what left you unsettled?
      • How would you describe the overall confidence level in your current modeling + response capabilities across senior leadership? Options: High — trusted and proven, Moderate — useful but with caveats, Low — causes repeated concern, Varies widely by function / team

      What If Estimates Aren’t Ready When the Board Calls?

      • Imagine a major event occurs and you don’t have timely loss estimates for the executive call—what immediate business consequences would you expect?
      • How quickly do you currently need preliminary, interim, and final loss estimates during an active event? Options: Preliminary: <1 hour / Interim: <4 hours / Final: days, Preliminary: 1–4 hours / Interim: 4–12 hours / Final: days, Preliminary: same day / Interim: daily / Final: weeks, Our requirements vary by stakeholder / event
      • Which stakeholders insist on the fastest estimates during an event? Options: Board / Executives, CFO / Finance, Reinsurance / Treaties team, Underwriting, Actuarial, Claims Operations, Regulatory / Compliance
      • When your estimates change materially during an event, how do you communicate updates and who must sign off on major revisions?
      • In the last three years, how often have delayed or inaccurate estimates caused material business problems (missed reinsurance notices, mispriced renewals, regulatory issues)? Options: Never, 1–2 times, 3–5 times, More than 5 times, Unsure

      Where Are Your Models Hiding Surprises?

      • Which parts of your modeling pipeline do you suspect are most likely to produce silent failures or misleading outputs?
      • Please list the primary data sources and integrations your catastrophe models rely on (policy systems, exposure feeds, third-party hazard data, claims feeds, etc.).
      • How often do you experience data gaps, mapping mismatches, or stale feeds that meaningfully degrade model accuracy? Options: Daily, Weekly, Monthly, Rarely, Never, Unsure
      • Which model components are you least comfortable explaining to non-technical stakeholders? Options: Hazard modeling, Vulnerability / fragility curves, Exposure mapping & geocoding, Aggregation / secondary uncertainty, Scenario selection & assumptions, Other
      • Describe a real example where a model surprise occurred—what was missed, how was it discovered, and what was the impact?
      • How long has this class of model surprise been recurring, and what fixes have you attempted so far?

      Who Pulls the Levers When Things Escalate?

      • If an active event requires external response activation, who ultimately decides to deploy adjusters, drones, or satellite services—and on what criteria?
      • Which roles are formally included in your incident command structure for catastrophes? Options: CRO, Head of Catastrophe Risk, Chief Actuary, CIO, VP Claims, Head of Cat Claims Ops, COO, Chief Legal, Head of Underwriting, Other
      • Are your escalation paths and runbooks documented and rehearsed, or are decisions typically made ad hoc? Options: Fully documented and rehearsed, Partially documented, Ad hoc with some playbooks, Not documented
      • What political, budgetary, or operational frictions typically surface when you try to scale external response resources?
      • When ownership is unclear, which factors most commonly slow decision-making: approval gates, budget authorization, data access, vendor onboarding, or something else? Options: Approval gates, Budget authorization, Data access / mapping, Vendor onboarding, Legal / compliance review, Other
      • Has ambiguity in governance ever led to delayed payments, slower customer resolution, or reputational impact? Please share an example if comfortable.

      What Would Real-Time Confidence Look Like for You?

      • What would it take for you to trust a real-time loss estimate enough to act on it without waiting for manual executive sign-off?
      • What accuracy tolerances would you require for preliminary, interim, and final estimates? Options: Preliminary ± >20% / Interim ±10–20% / Final ± <10%, Preliminary ±10–20% / Interim ±5–10% / Final ± <5%, Preliminary ±5–10% / Interim ±3–5% / Final ± <3%, Unsure / varies by use case
      • What maximum latency is acceptable for preliminary estimates to be operationally useful for triage and reserving actions? Options: <1 hour, 1–4 hours, 4–12 hours, Same day, Daily
      • Which three KPIs would you prioritize during an active event to decide whether to scale response or trigger reinsurance notifications? Options: Estimate accuracy, Estimate latency, Adjuster fill-rate, Claims intake volume, Time to first notice, Customer satisfaction, Reinsurance notification timeliness, Other
      • How would you want to validate model outputs in real time—automated cross-checks, sample-field validation, external benchmarks, or something else? Options: Automated backchecks, Field validation (adjuster/drones), Satellite imagery correlation, External independent benchmarks, Ad hoc manual validation, Other
      • Would automated alerts tied to KPI thresholds and confidence bands be useful—and which teams should receive them? Options: Yes — Risk / Actuarial, Yes — Claims Operations, Yes — CRO / Execs, Yes — Underwriting, No, Unsure

      What Operational Gaps Would Break a Deployment on Day One?

      • What infrastructure or access gaps today would prevent you from standing up a tested integration in a sandbox or production environment?
      • Which of the following are already provisioned and tested in your environment? Options: Policy feed / exposure feed, Claims feed, Ingest API keys / credentials, Sandbox / test environment, Production API access, Adjuster roster / dispatch integration, Drone / satellite vendor integration, Command center contact lines
      • Who is responsible for provisioning and testing the integrations and feeds: internal IT, data engineering, claims ops, vendor, or a combination? Options: Internal IT, Data Engineering, Claims Operations, Vendor / Supplier, Shared ownership, Unsure
      • How often do you run full drills or runbook rehearsals that include model refresh, estimate publication, and response dispatch? Options: Monthly, Quarterly, Twice a year, Annually, Never
      • Which technical or operational failure modes worry you most during a cutover (select up to three)? Options: Data mapping mismatches, API authentication / throttling, Latency under load, Incomplete adjuster coverage, Vendor coordination failures, Regulatory / legal constraints, Other
      • How confident are you in your current test coverage and rollback plans on a scale from 0 (no confidence) to 10 (fully confident)? Options: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10

      If We Partnered, What Would Success Actually Look Like?

      • If we worked together and an event occurred tomorrow, what would success look like 90 days later—and what single metric would you point to as evidence?
      • Which commercial or governance commitments would make you comfortable moving from a pilot to full production? Options: Latency SLA, Accuracy SLA, Data-sharing & security agreement, Escalation path / governance, Pricing / commercial model, Dedicated support and training, Joint roadmap & backlog, Other
      • Which internal stakeholders would need to be convinced for a mutual commit, and what evidence would satisfy each (briefly list role → ask / metric)?
      • What format and cadence for post-event retrospectives would be most useful to your team (attendees, outputs, and follow-up expectations)?
      • How would you prefer ongoing prioritization for model improvements and response enhancements to be managed? Options: Joint governance board, Quarterly roadmap reviews, Vendor-driven backlog with tickets, Ad-hoc operational sprints, Other
      • Realistically, what are the next steps you expect from us after this discovery conversation? Options: Detailed technical proposal, Executive briefing, Pilot integration / sandbox test, Contract & commercial draft, Technical deep-dive workshop, Other

    2. Deployment Enablement

      Schedule cutovers, drills, runbook rehearsals, and task ownership to operationalize event activation and claims response.

    3. Validation Checklist

      Verify end-to-end accuracy of loss estimates, integration test results, response dispatch workflows, and KPI measurement prior to handoff.

      Validation Questions

      Quick Check: Where Are We Right Now?

      • Who will be our primary point(s) of contact for validation, handoff, and escalation? Options: CRO, Head of Cat Risk, Chief Actuary, CIO, VP Claims, Head of Cat Claims Ops, Integration Lead, Other
      • Walk us through your current end-to-end validation process for loss estimates and integrations—what steps happen today, who runs them, and when?
      • Which of these systems does our platform already integrate with in your environment? Options: Policy administration system, Claims system, Reinsurance accounting, Data warehouse / BI, Geocoding/GIS, Weather/event feeds, API gateway / ESB, None yet, Other
      • How often do you run full end-to-end validation or integration tests outside of real events? Options: Daily, Weekly, Monthly, Quarterly, Only before major changes, Ad-hoc / as needed
      • When your last full validation ran, what succeeded and where did you still see gaps?

      What If Your Loss Numbers Were Wrong?

      • If our pre-event or live-event loss estimates were later shown to be materially off, who inside and outside your organization would face the biggest consequences? Options: CRO, Chief Actuary, CFO, Board, Reinsurers, Claims leadership, Regulators, Other
      • How do you define 'materially off' for your purposes—by percentage error, reserve variance, market impact, or something else? Options: ±5% error, ±10% error, ±20% error, Reserve impact threshold, Defined by stakeholder impact, Other
      • Tell us about a past incident where an estimate missed materially—what happened, why did it occur, and what were the downstream impacts?
      • For pricing, reserving, and live-event briefing needs, what accuracy tolerances would you require for initial and revised estimates?
      • How tolerant are your external stakeholders (board, reinsurers) to revisions after the first public estimate? Options: Very tolerant (expect refinements), Somewhat tolerant, Low tolerance — expect high stability, Depends on magnitude and transparency

      When Integrations Break on Day One

      • When a schema change or missing feed happens in a live event, how quickly do your integrations typically fail or degrade? Options: Immediately / within minutes, Within hours, Over a day, We don't know / no monitoring
      • Which data feeds are mission-critical for accurate real-time loss estimation and dispatch? Options: Exposure / policy details, Claims intake, Geocoded addresses, Live weather feeds, Satellite imagery, Drone feeds, Adjuster rosters, Reinsurance treaties and coverage terms, Other
      • Describe your ideal error-detection and alerting cadence during a live event—what counts as an alert and who must be notified?
      • Which uptime and recovery SLAs do you require for critical feeds and APIs during an event? Options: 99.99% (near-continuous), 99.9%, 99%, 95%, We don't have formal SLAs yet
      • Who on your team owns feed troubleshooting, and how do escalations to vendor/partner teams need to be structured?
      • Are you able to run replay tests using historical events to validate integrations and outputs? If yes, when was the last replay and what surfaced? Options: Yes—recently (past 6 months), Yes—but not recently, Planned, No

      Dispatch Under Pressure: Can We Move Fast Enough?

      • If claims surge to 5x your normal volumes within 48 hours, how confident are you that dispatch workflows and external adjuster rosters will scale to meet service targets? Options: Very confident, Somewhat confident, Low confidence, Not confident / major gaps
      • What is your current maximum claims-per-adjuster expectation during a major catastrophe (rough bands are fine)? Options: <50, 50–100, 100–200, >200, Unknown / varies widely
      • Which response resources do you expect us to supply or coordinate during activation? Options: Licensed adjusters, Drone-based aerial assessments, Satellite imagery analysis, Mobile claims processing units, Local contractor networks, Triage call centers, Other
      • Walk us through your onboarding, licensing, and credentialing requirements for external adjusters—what usually slows this down?
      • What parts of the dispatch process tend to become bottlenecks (triage, travel/logistics, data handoff, report consolidation, payables)? Options: Triage / prioritization, Travel and logistics, Data handoff to claims teams, Adjuster reporting consolidation, Payment/administration, Other
      • How would you prefer we demonstrate response readiness—tabletop exercises, live drills, joint deployments, or detailed after-action reports? Options: Tabletop exercises, Full drills (field), Live-seasoned deployments, Detailed after-action reports, Combination of the above

      Who Signs Off on 'Good Enough'?

      • Who in your organization would have final veto authority over a validation pass—and what evidence would make them approve it? Options: Chief Actuary, CRO, CIO, VP Claims, Head of Cat Risk, Legal/Compliance, External auditor, Other
      • What concrete acceptance criteria must be met for model accuracy, run-time performance, and report delivery before handoff?
      • What report formats and cadences satisfy your audit, board, and regulator expectations (e.g., executive PDF, drillable dashboard, raw CSV/API)? Options: Executive PDF + appendix, Interactive dashboard, API endpoints / raw CSVs, Daily email briefings, Custom regulator packages, Other
      • Do you require independent third-party validation, internal model governance sign-off, or both before accepting our outputs? Options: Internal governance only, Independent third-party validation, Both, Unsure / want recommendations
      • Which KPIs should appear on the operational validation dashboard for handoff (pick the most important)? Options: Loss estimate error / bias, Model runtime per scenario, Data latency by feed, Claims triage time, Adjuster assignment time, API success rate, Number of failed automated checks
      • How often will governance formally review validation results once the system is live (during season and off-season)? Options: Daily (during event), Twice daily, Weekly, Monthly, Quarterly, On-demand / as-needed

      When Things Go Off Script

      • What is your appetite for automated fallbacks if a primary data feed or model component fails during an event—fully automated, human-reviewed, or no automation? Options: Fully automated fallback, Human-reviewed fallback, No automatic fallback (manual only), Depends on the component
      • Which fallback mechanisms would you like us to support or build (statistical imputation, reduced-scope models, manual override queues, cached results)? Options: Statistical imputation, Reduced-scope / regional models, Manual override queue, Cached last-known-good results, Graceful degradation visuals, Other
      • Describe a contingency that previously either saved an event response or failed to help—what did you learn?
      • How should responsibility be split between our operations and yours during a fallback (we lead, you lead, shared with predefined triggers)? Options: We lead with customer oversight, You lead with vendor support, Shared with pre-defined triggers, Case-by-case
      • What contractual remedies, SLA credits, or escalation expectations would you expect if a critical validation check fails during an event?
      • How quickly do you expect a root cause analysis after a failure, and what level of detail do you need in that RCA? Options: Initial 24h summary, detailed in 72h, 72h detailed RCA, 2 weeks full investigation, As requested / case-by-case

      What Would Calm the Board (and Your Reinsurers)?

      • If you had one trusted artifact to present to the board and reinsurers within 12 hours of landfall, what must it show and why?
      • Which audiences need tailored versions of the same brief (CRO, CFO, Board, Reinsurers, Regulators), and what information is most critical for each? Options: CRO, CFO, Board, Reinsurers, Regulators, Claims leadership, Investors, Other
      • Which metrics or visuals should be prioritized in initial communications (total insured loss, geographic heatmap, top exposure clusters, reserve ranges, confidence intervals)? Options: Total insured loss, Geographic heatmap, Top-10 ZIPs/regions, Reserve range with CI, Claims volume projections, Model confidence bands
      • How comfortable are you with publishing initial estimates accompanied by explicit confidence intervals and uncertainty narratives? Options: Very comfortable, Somewhat comfortable, Prefer point estimates only, Depends on audience
      • What cadence of update bulletins do external stakeholders expect in the first 72 hours (e.g., hourly, every 4 hours, twice daily)? Options: Hourly, Every 4 hours, Twice daily, Daily, As material changes occur
      • Do you require a signed attestation from our model operations team for any published numbers to external audiences? Options: Yes, for board/reinsurer briefs, Only for regulators, No, Sometimes / scenario-based

      Ready to Sign Off?

      • What would make you say 'yes' to operational handoff today—be specific about artifacts, tests, and demonstrations?
      • Which final validation artifacts are must-haves before accepting handoff? Options: Full integration test logs, Passed end-to-end scenario runs, Signed runbooks and SOPs, Drill reports / exercise results, KPI dashboard with baselines, Acceptance signoff document
      • Who must sign the final acceptance, in what order, and are there any procurement or legal approvals we should schedule in advance? Options: Chief Actuary, CRO, CIO, VP Claims, Legal/Compliance, Procurement, Other
      • After passing validation, what timeline do you expect between signoff and being ready to operate in a live event? Options: Same day, 1–3 days, 1 week, 2+ weeks, Depends on outstanding items
      • How should we structure a shared backlog for post-handoff improvements (tool choice and ownership model)? Options: Shared Jira board with tags, CustomerNode task board, Weekly triage calls + email tracker, Shared spreadsheet, Other
      • List the top three risks you want on the joint risk register before handoff and any immediate mitigations you expect.

  7. Success

    Review outcomes against success signals, run post-event retrospectives, and maintain a shared backlog for issues and enhancements.

    Success Reviews

    • Post-Event Outcomes Review (Executive & Ops)
    • Technical Retrospective & Root Cause Analysis
    • Claims & Response Operations Retrospective
    • Shared Backlog Prioritization & Roadmap Alignment (Customer + Product)
    • SLA, Commercial & Governance Review

    Issues & Enhancements

    • Define communication cadence and metrics to track progress back to the customer.
    • Operational Timeline & Throughput Review
    • Document operational failures and wins with clear owners for runbook updates and staffing changes.
    • Agree on immediate tactical changes to improve adjuster dispatch and assessment delivery times.
    • Schedule follow-up drills and define success criteria to validate operational fixes.
    • Update runbooks with agreed edits and circulate a 'versioned' runbook to all operational teams.
    • Re-certify/augment adjuster rosters for surge response and identify training needs.
    • Schedule a full-field drill with vendors and customers within the next quarter to validate changes.
    • Backlog Framing & Prioritization Criteria
    • Produce a prioritized, time-bound shared backlog with owners and acceptance criteria for each item.
    • Place high-impact fixes into the product/ops roadmap with agreed release targets.
    • Opening & Meeting Objectives
    • Publish the shared backlog with priority, owner, target date, and customer acceptance criteria.
    • Create pilot test plans for top 3 customer-impact items and schedule acceptance windows.
    • Set up a fortnightly steering update with agreed KPIs to show incremental progress.
    • SLA Performance Summary
    • Determine and document any commercial remediation or credits arising from SLA breaches.
    • Agree on governance and escalation changes to prevent recurrence and speed future activations.
    • Set a clear path and timeline for any contract amendments and customer sign-off.
    • Draft and circulate agreed commercial remediation language or credit memo for customer approval.
    • Update the governance charter and escalation matrix and obtain executive sign-off.
    • Publish a short FAQ for internal and customer-facing teams explaining remediation and next steps.
    • Create a validated, shared account of how outcomes compared to the agreed success signals.
    • Authorize immediate remediation actions for high-impact gaps and assign accountable owners.
    • Confirm follow-up meetings and deliverables required for the retrospective and backlog work.
    • Produce a one-page signed outcomes statement (what met target, what missed, material impact) and circulate to stakeholders.
    • Assign owners and due dates for immediate remediations (data fixes, runbook changes, claims surge augmentation).
    • Schedule technical retrospective and backlog prioritization workshops within 7 business days.
    • Framing: Current State, Consequence, Desired Future State
    • Identify root causes for each major technical failure and agree prioritized corrective actions with owners.
    • Define measurable validation steps and acceptance criteria to prove fixes restore the desired future state.
    • Establish timeline for technical fixes and re-test cadence.
    • Create a prioritized technical fixes log with severity, owner, target date, and verification method.
    • Provision a test environment and schedule validation runs for each fix with required datasets.
    • Implement monitoring/alerting adjustments to surface the same failure modes in future events.
    • Contractual Implications & Remediation Options
    • Review & Triage Top Items
    • Summary of Event Timeline & Decisions
    • Walkthrough of Failure Modes (Diagnosis)
    • Field Ops Case Studies
    • Success Signals vs Measured Outcomes
    • Governance & Escalation Pathway Revisions
    • Resource Capacity & Roster Effectiveness
    • Roadmap Placement & Release Windows
    • Impact Quantification by Failure Mode
    • Financial & Operational Impact
    • Problem-Solving: Corrective & Preventive Actions (Proof)
    • Service Credits / Commercial Adjustments (Decision)
    • Customer Acceptance & Pilot Plans
    • Runbook & Drill Effectiveness
    • Agreement on Contract Amendments & Next Steps
    • Gaps & Immediate Remediations
    • Operational Improvements & Tactical Decisions
    • Validation Plan & Acceptance Criteria
    • Governance & Communication Plan
    • Decisions & Next Steps
Explore more templates Use this template
First-Party AI ™

1-2 minutes please — Your AI agent is working

First-Party AI™ can make mistakes. Always check important information.