Fragmented context
Data, assumptions, methods, commentary, and approvals live in separate places.
Governed actuarial intelligence for reserve close.
ActuitMCP connects insurance data, actuarial methods, expert judgment, reporting, and evidence in one controlled workflow. Use it as part of the full Facio platform or connect it to your existing environment as a governed MCP utility.
Early demo access will open first to selected teams in the United States.
Deterministic calculations · Human approval · Source-to-report lineage
Commercial Auto · June 2026
Gross reserve analysis
Paid-loss triangle2026-Q2-v3
Paid CL
$18.4m
Incurred CL
$19.1m
BF
$18.8m
Approved AI client
Compare paid CL, incurred CL, and BF. Keep the large-loss scenario separate.
Governed workflow
Three deterministic methods are ready. No selection has been approved.
Artifact status
Draft · Review required
Illustrative product experience
P&C Reserve Close
Built around practical reserving workflows
Web + MCP
The same governed capabilities through two interfaces
Multi-format Output
HTML, PDF, XLSX, PPTX, JSON, and evidence packages
Human-Owned Judgment
AI drafts and explains; authorized actuaries decide
The calculation is only one part of the work.
Reserve analysis is often spread across source extracts, spreadsheets, actuarial software, notebooks, review emails, committee decks, and audit folders. Reproducing how the result was reached can still take days.
ActuitMCP is being built to connect the complete path: data checks, triangles, methods, judgment, scenarios, reporting, review, and evidence.
Repeated reconstruction
Each close rebuilds mappings, workpapers, slides, and audit support.
Weak traceability
A material figure may be difficult to follow back to its exact source snapshot, method, and selection.
Built into Facio. Available beyond it.
Actuit runs on the same governed actuarial foundation in both delivery models. Choose the experience that fits your insurance stack.
Actuarial intelligence inside the insurance operating system.
For teams using the full Facio solution, Actuit becomes the actuarial layer over connected policy, premium, claims, document, and reporting state. Fewer extracts, fewer handoffs, and one governed operating history.
- Native access to approved Facio operational data
- Connected policy, premium, claims, and reporting context
- Shared identity, permissions, audit, and workflow controls
- Approved outputs available to downstream operations
Connect governed actuarial workflows to the systems you already use.
Use ActuitMCP alongside existing claims, policy, finance, data, and AI environments. Bring governed data in, run controlled actuarial workflows, and return approved results through web, API, and MCP interfaces.
- Secure files and supported data connectors
- Governed MCP tools for approved AI clients
- Independent workspace for review and approval
- Exportable reports, data, and evidence packages
The utility does not require the Facio policy administration system. The full platform provides the deepest operational continuity through the same governed insurance state.
From source data to a reviewable reserve decision.
Designed around the complete reserve-close chain, not an isolated calculator or generic chatbot.
01
Bring data in without losing its history.
Connect files, object storage, databases, SharePoint, or Google Drive. Actuit profiles the source, proposes canonical mappings, preserves the original, and freezes approved versions for calculation.
- Schema and statistical profiling
- Mapping confidence and review
- Immutable source snapshots
- Multilingual input with governed English normalization
02
Know whether the data is ready before the method runs.
Validate completeness, dates, identities, duplicates, source totals, prior-period changes, and ledger reconciliation. Material exceptions remain visible and can block release.
- Paid, case, and incurred checks
- Triangle-to-source reconciliation
- Prior-valuation change detection
- Materiality-aware exceptions
03
Build triangles that can explain every cell.
Create paid, incurred, case, count, premium, and exposure triangles by segment, basis, grain, currency, and large-loss treatment, with cell-to-source drill-down.
- Paid and incurred development
- Claim-count and exposure views
- Calendar-period diagnostics
- Cell-to-source drill-down
04
Compare methods without hiding the judgment.
Run transparent deterministic methods such as Chain Ladder, Bornhuetter-Ferguson, Cape Cod, Expected Loss Ratio, and approved blends before an actuary records a selection.
- Versioned calculation engines
- Explicit assumptions and exclusions
- Development and tail comparison
- Human-owned selection rationale
05
See what changed, why it changed, and what happens next.
Compare valuation periods and controlled scenarios for factors, tails, method weights, large-loss treatment, and expected loss ratios. Preserve every alternative to the baseline.
- Actual versus expected
- Quarter-over-quarter movement
- Large-loss treatment
- Scenario comparison and impact
06
Generate the report and its evidence from the same result.
Produce branded executive, actuarial, data-quality, scenario, and governance outputs from one approved payload, keeping figures consistent across every required format.
- Consistent figures across formats
- Source-to-report lineage
- Review and approval status
- Immutable artifact checksums
A reserve-close workflow that stays connected.
- 01
Connect
Register or upload approved source data.
- 02
Validate
Map, reconcile, and resolve material exceptions.
- 03
Analyze
Build triangles, factors, tails, and method results.
- 04
Decide
Record expert judgment, selections, and scenarios.
- 05
Review
Complete peer review, validation, and approval.
- 06
Release
Generate reports, data outputs, and evidence.
Every step records the data version, method version, assumptions, actor, status, and resulting artifact.
Work visually. Ask naturally. Govern both the same way.
A conversation can prepare an analysis, but the underlying data checks, calculations, permissions, review, and approval remain deterministic and governed.
Inspect and approve the work.
- Review mappings and exceptions
- Inspect triangles and diagnostics
- Compare methods and scenarios
- Record assumptions and judgment
- Review, approve, and release artifacts
Orchestrate from an approved AI client.
- Discover actuarial tools and templates
- Create valuations, triangles, methods, and scenarios
- Ask for explanations backed by stored calculations
- Generate draft reports and evidence
- Submit governed work for human review
The control boundary: AI can plan, explain, translate, and draft. It cannot create authoritative reserve figures, approve a selection, bypass validation, or release an artifact.
One request. A complete, reviewable workflow.
User request
Run the June 2026 gross reserve analysis for Commercial Auto. Build paid and incurred annual triangles, keep the catastrophe event separate, compare Chain Ladder, Bornhuetter-Ferguson, and Cape Cod, and prepare the close pack.
Actuit · 01
The snapshot is reconciled. One material mapping exception remains open, so release is blocked. Draft analysis can continue.
Actuit · 02
Three deterministic methods are complete. The largest variance is in accident year 2025.
Actuit · 03
A comparison and draft selection pack are ready for actuarial review. No selection has been approved.
Method comparison · USD millions
Commercial Auto gross reserve
| Method | Ultimate | IBNR | vs. selected |
|---|---|---|---|
| Paid Chain Ladder | $67.8 | $18.4 | −2.1% |
| Incurred Chain Ladder | $68.5 | $19.1 | +1.6% |
| Bornhuetter-Ferguson | $68.2 | $18.8 | — |
Reserve movement+$1.7m QoQ
Open
Paid
Case
Assump.
Close
Close packDraft · Review Required
Reports that do not need to be reconciled back to each other.
Each format is rendered from the same immutable result payload. The narrative may explain a figure, but it does not recalculate it.
Interactive workspace
Filters, drill-down, review, annotations, and lineage
Executive PDF
Branded close summary with assumptions, movement, and status
XLSX workpaper
Structured actuarial tables with protected provenance
PPTX pack
Board and committee-ready slides linked to approved figures
JSON and Parquet
Versioned machine-readable output for downstream systems
Evidence package
Checksums, source manifests, method versions, approvals, and artifacts
Actuit is being designed around a governed catalog of 295 actuarial, data-quality, reporting, scenario, and evidence definitions. Initial demo access will focus on the Reserve Close Core.
Automation accelerates the work. It does not own the decision.
Designed for environments where a result must be explainable, reviewable, and reproducible after the conversation ends.
Deterministic calculations
Material figures come from versioned actuarial engines, not language-model arithmetic.
Human approval
AI can propose and explain. Authorized professionals select, review, and approve.
Immutable source versions
Every run points to exact data, assumption, method, and template versions.
Source-to-report lineage
Material figures resolve through the calculation graph to contributing source records.
Controlled release
Draft, validated, reviewed, approved, and released are distinct states.
Regional data controls
Initial demo access is planned for selected US teams. Production options are confirmed during sales and security review.
ActuitMCP supports actuarial analysis and governance. It does not replace professional actuarial judgment, provide autonomous actuarial sign-off, or constitute regulatory approval.
Enterprise-grade actuarial workflow without rebuilding an enterprise stack.
Consulting actuaries
Standardize repeatable client analysis, branded reporting, review, and evidence while preserving engagement-specific judgment.
MGAs and program administrators
Add reserve intelligence and professional governance without assembling a new actuarial engineering platform.
P&C insurers and reinsurers
Connect portfolio data, reserve methods, movement analysis, and approval across segments and valuation periods.
Insurance technology teams
Integrate governed actuarial workflows through web, API, and MCP interfaces while retaining existing source systems.
Be among the first teams to evaluate ActuitMCP.
We are preparing the first ActuitMCP demo program for selected insurance and actuarial teams in the United States. Tell us about your reserve-close workflow and how you would want to use Actuit.
No production insurance data
Multilingual source input
Human-owned approval
What to know before the demo.
What is ActuitMCP?
ActuitMCP is a governed actuarial intelligence service for P&C reserve-close workflows. It connects approved data to validation, triangles, deterministic methods, actuarial selections, scenarios, reports, and evidence through a web workspace and MCP-compatible interfaces.
Do I need the full Facio platform?
No. ActuitMCP is planned as both a native capability inside the full Facio platform and a focused utility that can work alongside approved external data and insurance systems. The full platform provides deeper operational continuity.
What does MCP mean here?
MCP is a standard way for approved AI clients to discover and call governed tools. It lets a user request workflows conversationally while permissions, data checks, calculations, review, and approvals remain controlled by Actuit.
Does the AI calculate the reserve?
No. Material actuarial figures are produced by versioned calculation engines. AI may interpret a request, orchestrate approved tools, explain stored results, translate source text, and draft commentary. It cannot approve a reserve or replace professional judgment.
Which reserving methods are planned?
The Reserve Close Core is designed around paid and incurred Chain Ladder, Bornhuetter-Ferguson, Cape Cod, Expected Loss Ratio, Average Cost, and approved method blends, together with development and tail selection.
What data can ActuitMCP accept?
The planned service supports structured claim, policy, premium, exposure, finance, and prior-valuation data through approved files and connectors. Data must pass mapping, quality, reconciliation, authorization, and regional-policy controls.
Can it accept non-English source data?
Yes. ActuitMCP is designed to accept multilingual source material while preserving the original evidence. The application, reports, and approved outputs are produced in English and left-to-right format.
When will the demo be available?
Facio is preparing an initial demo program for selected US-based teams. A public release date has not been announced. Joining the waitlist registers interest but does not guarantee timing or access.
Is ActuitMCP a regulatory filing or actuarial sign-off service?
No. It supports controlled analysis, documentation, review, and evidence. Regulatory submission and professional actuarial sign-off remain the responsibility of authorized people and organizations.
Make the reserve close easier to run — and easier to defend.
Join the US demo waitlist to learn how ActuitMCP could connect your data, actuarial workflow, reporting, and evidence.