Prior Experience

Pilot program for captive managers and actuarial teams

From messy loss runs to review-ready captive analysis.

Prior Experience helps captive managers and actuarial teams normalize inconsistent claim data, preserve assumptions, run repeatable reserve workflows, and produce controlled work products without adding another spreadsheet process.

Book a Captive Loss-Run Clinic View Synthetic Sample

Synthetic or redacted data is welcome. Production claim data is not required for an initial review.

Illustrative synthetic workflow — a simplified representation of the ingest–validate–analyze sequence, not screenshots of the live application.

Captive work often breaks between the source file and the final opinion.

Inconsistent inputs

Different TPAs use different headers, dates, claim identifiers, status labels, valuation practices, and financial definitions. The same captive can receive three loss runs that disagree about what "incurred" means.

Repeated manual work

Teams repeatedly clean files, rebuild spreadsheet logic, transfer assumptions, and reconcile changing datasets — every quarter, for every entity, from scratch.

Review friction

Decisions, exceptions, and assumptions can be difficult to trace during peer review, board reporting, audit, or regulatory review. The reasoning lives in someone's spreadsheet, or in someone's head.

One controlled workflow from source file to work product.

Each stage preserves what came before it. Source values, mappings, exceptions, and assumptions stay attached to the analysis they produced.

Ingest

Accept representative PDF, XLSX, and CSV loss runs as they actually arrive from TPAs.

Validate

Map fields, preserve source values, flag inconsistencies, and reconcile paid, case, and incurred totals.

Analyze

Create development views and run repeatable reserve calculations on the normalized claim history.

Review

Record assumptions, selections, exceptions, notes, and approvals where a reviewer can find them.

Deliver

Generate controlled PDF, Excel, and structured outputs with version and run identifiers.

Professional boundary. Prior Experience is designed to support professional analysis, review, and documentation. It does not replace required actuarial, management, regulatory, or audit judgment.

Inspect the work product, not just the feature list.

Every sample below comes from one internally coherent synthetic case. The arithmetic reconciles, and the assumptions remain visible for review.

Page 2 of the draft synthetic reserve analysis: a table of accident-year ultimate loss estimates comparing paid losses to date, the credibility-model indicated ultimate, the chain-ladder ultimate, and the selected ultimate, above a chart of paid losses with indicated-ultimate uncertainty intervals for accident years 2021 through 2026. The page carries a DRAFT — pending review watermark.

Page 2 of the draft reserve analysis, generated by the pilot application from six synthetic TPA loss-run snapshots (run 7347e4dd).

Synthetic pilot output Draft — pending review

Reserve analysis by accident year

Compares modeled ultimate losses with a permanent chain-ladder reference and shows the uncertainty range behind the selected estimate.

Every figure on the page traces back through the report: age-to-age factors computed from the development data in its appendix, chain-ladder ultimates that reconcile to the reference column, and stamped assumptions — tail factor, benchmark prior, credibility weight — recorded in the audit trail rather than implied.

See the full sample walkthrough →

Run 7347e4dd · eval 2026-12-31 · synthetic WC case

Synthetic pilot output

Loss-run validation and lineage

Six annual TPA snapshots reconciled claim-by-claim, each tied to its printed totals before any analysis ran. From the report's audit trail:

Loss-run snapshot lineage from run 7347e4dd
SnapshotEvaluationClaimsTie-out
792ed0a02021-12-3121tied
a623c10d2024-12-3182tied
67737bd02026-12-31119tied

Excerpt — three of six snapshots shown.

View validation samples →

Synthetic pilot output Draft — pending review

Reserve funding by confidence level

Modeled reserve-funding views at stated confidence levels — not premium recommendations. From the funding exhibit:

Reserve funding by confidence level from run 7347e4dd
LevelRequired funding
Expected (mean)$245K
75th percentile$409K
90th percentile$596K
Chain-ladder (reference)$833K

View the funding exhibit →

The synthetic case behind these samples

Scenario: six annual TPA loss-run snapshots (2021–2026) for one synthetic workers'-compensation captive, reconciled claim-by-claim across evaluation dates. Output: a reserve analysis with a permanent chain-ladder reference, a modeled uncertainty range, a funding-by-confidence-level exhibit, and an audit trail that stamps every assumption, data snapshot, and diagnostic onto the run.

Designed to reduce repeated cleanup and make the analysis easier to review, reproduce, and regenerate. Time and cost savings will be reported only after they are measured in real pilots.

What you leave the clinic with.

Data-quality assessment

A focused review of structural inconsistencies, missing fields, financial-definition questions, and reconciliation risks.

Proposed mapping and exception list

A practical view of how source fields could be standardized and which records would still require human review.

Workflow and pilot-fit assessment

A walkthrough of how the file could move from ingestion through reserve analysis, review, and controlled reporting.

The clinic is a product-discovery and workflow-review session. It does not produce an actuarial reserve opinion.

Sensitive-data handling should be clear before the first call.

Claim data deserves explicit answers, not a padlock icon. Every control is labeled with its real status.

Implemented

Synthetic or redacted data is accepted for initial evaluations. Production claim data is not required for the first clinic.

Implemented

The general contact form does not accept file uploads. Data-sharing instructions are provided separately after a clinic is booked.

Pilot

Data handling, retention, access, and deletion rules are documented and shared with every pilot participant before any file is exchanged.

Planned

Dedicated-instance, customer-cloud (AWS account or VPC), and customer-hosted deployment paths are under evaluation. The shared service is the intended standard offering.

Read the full data-handling note →

Start with one difficult loss run.

Bring a synthetic, redacted, or representative loss-run file. We will review its structure, identify ingestion and reserving challenges, and show how the workflow could be standardized.

The clinic is a product-discovery and workflow-review session. It is not an actuarial opinion, insurance recommendation, regulatory filing, or audit.

Read the Pilot Overview →

Book a Captive Loss-Run Clinic

Free, 15–30 minutes. Four quick fields; scheduling happens after your request is received.

Enter your name.

Enter a valid work email address.

Enter your company.

Select your role.

Do not send production claim data through this form. Synthetic or redacted examples can be discussed after the initial booking. Your details are used only to respond, schedule, and assess pilot fit — see the privacy notice.

Questions a careful reviewer should ask.

Who is Prior Experience built for?

Captive managers and consulting actuarial teams who work with recurring TPA loss runs and reserve analyses. Captive owners, finance teams, auditors, and directors benefit from the outputs, but the workflow is designed around the people who prepare and review the analysis.

Is this a replacement for an actuary?

No. Prior Experience prepares, validates, and documents the data and reserve analysis. Reserve opinions, actuarial sign-off, and professional judgment remain with qualified professionals. The product is designed to make their review easier, not to substitute for it.

What does the Captive Loss-Run Clinic include, and what does it cost?

The clinic is free and runs 15–30 minutes. It covers a structural review of one representative loss run, a data-quality assessment, a proposed field mapping, an exception list, workflow observations, a product walkthrough, and a pilot-fit assessment. Bring synthetic, redacted, or representative material — production claim data is not required and should not be sent. It is not an actuarial opinion or audit.

Is production claim data required?

No. Synthetic or redacted data is welcome, and the first clinic is designed to run without production data.

Which loss-run formats and TPA variations can be reviewed?

PDF, XLSX, and CSV loss runs. Handling inconsistent TPA formats is the core of the product: mapping different headers, dates, identifiers, status labels, and financial definitions onto one normalized claim history while preserving original source values. The clinic exists precisely to look at your representative formats.

Can assumptions, calculations, exceptions, and versions be reviewed?

Yes. Assumptions, selections, exceptions, and review notes are recorded alongside the analysis, and every generated work product carries a run identifier and version, so reports can be regenerated from the same inputs and differences between runs can be explained by what changed.

How are data handling, hosting, and deployment options addressed?

The data-handling note covers collection, processing and storage, access, retention, deletion, backups, data and work-product ownership, subprocessors, and deployment models — each labeled Implemented, Pilot, Planned, Customer-dependent, or Not available. The pilot page distinguishes current from planned scope.

More questions — product stage, what is published, and what is not yet available →

Start with one difficult loss run.

Book a Captive Loss-Run Clinic

Synthetic or redacted data is welcome. Production claim data is not required for an initial review.