Four products. One analytics engine.
PolicyLens serves commercial insurance. CLEAR, Proposal Intelligence, and TRACE serve federal audit compliance, procurement, and testable requirements. All turn dense records into fast, cited, decision-ready analysis.
PolicyLens
PrototypeCommercial · Insurance Intelligence
Decision-grade policy comparison, analytics, and recommendations — authored by AI.
See detailsCLEAR
PrototypeFederal · Audit & Compliance Analytics
Crosswalk DoD cost-accounting requirements to commercial equivalents — and see what audit reliance is worth.
See detailsProposal Intelligence
BetaFederal · Procurement Intelligence
Find the right opportunities — and the whole landscape behind them.
See detailsTRACE
PrototypeFederal · Requirements & Autonomy Analytics
Turn aspirational autonomy language into requirements you can test.
See detailsPARE
PrototypeFederal · Acquisition & FAR Reform
Pare the FAR back to what statute actually requires — clause by clause.
See detailsCEDAR
PrototypeFederal · Acquisition & Commercial-First Reform
The defense-unique answer is now the one that has to be justified.
See detailsPolicyLens
PrototypePolicyLens ingests carrier illustrations, actuarial memos, and regulatory filings, extracts the numbers that matter, and turns them into side-by-side comparisons, cited analytics, and client-ready recommendations — in seconds.
How PolicyLens works
Structured extraction
From PDFs to typed, comparable data.
An LLM pipeline reads carrier illustrations and actuarial memos, pre-scanning the full document so labeled values aren't lost to the truncation window.
See detailsSide-by-side comparison
Compare any number of carriers at once.
N-policy comparison across cost efficiency, IRR, and stress scenarios — best and worst values highlighted per metric, with guarantees decoded from the actuarial memo.
See detailsAnalytical Q&A
Ask questions — get numbers, cited.
Retrieval-augmented Q&A over filings and illustrations routes numerical questions through statistical analysis and text questions through grounded retrieval, always citing its sources.
See detailsClient recommendations
A defensible recommendation in one click.
Scores policies on normalized metrics, picks a carrier, and generates a grounded recommendation — with client and advisor cover pages — as PowerPoint, PDF, or Word.
See detailsStructured extraction
An LLM pipeline reads carrier illustrations and actuarial memos, pre-scanning the full document so labeled values aren't lost to the truncation window.
- PDF & Excel ingestion
- Table extraction
- Full-text pre-scan
- Schema-driven, typed output
Side-by-side comparison
N-policy comparison across cost efficiency, IRR, and stress scenarios — best and worst values highlighted per metric, with guarantees decoded from the actuarial memo.
- Cost efficiency & IRR analysis
- Stress-scenario modeling
- 16 guaranteed provisions decoded
- Net illustrated yield ranges
Analytical Q&A
Retrieval-augmented Q&A over filings and illustrations routes numerical questions through statistical analysis and text questions through grounded retrieval, always citing its sources.
- FAISS grounded retrieval
- Statistics & trend detection
- SERFF filing analysis
- Source-cited answers
Client recommendations
Scores policies on normalized metrics, picks a carrier, and generates a grounded recommendation — with client and advisor cover pages — as PowerPoint, PDF, or Word.
- Data-driven carrier scoring
- Grounded, factual narrative
- Client & advisor cover pages
- PPTX · PDF · DOCX export
CLEAR
PrototypeCLEAR (Commercial Leverage & Equivalence for Audit Reliance) maps DoD-unique cost-accounting and business-system requirements (CAS, DFARS) to their commercial equivalents (GAAP, SOX 404(b), COSO), then derives — per supplier — what DCAA can rely on instead of re-auditing. Built air-gappable from day one, with every verdict agent-proposed and reviewer-approved.
How CLEAR works
Requirement crosswalk
Federal requirements, mapped to commercial standards.
A retrieve-then-reason agent maps DoD-unique cost-accounting and business-system requirements to their nearest commercial equivalents, streaming its rationale live so reviewers can follow the reasoning, not just the verdict.
See detailsSupplier reliance assessment
What can DCAA rely on instead of re-auditing?
For a given supplier, CLEAR derives a reliance map across every requirement — rely, retain, targeted test, or incremental — with reviewer overrides captured alongside the agent's proposal.
See detailsReliance-dividend analytics
Turn audit reliance into dollars, transparently.
Reliance dispositions convert into audit-hour and dollar savings using editable assumption drivers — audit hours, DCAA rate, contractor absorption — so every figure is auditable, not asserted. Rolls up into a portfolio view across suppliers and program concentration.
See detailsReviewer audit trail
Every determination, traced and defensible.
Every agent-proposed verdict and reviewer determination is logged to an audit trail, and the deployment itself — commercial cloud, GovCloud, or fully air-gapped enclave — is a configuration choice, not a rewrite.
See detailsRequirement crosswalk
A retrieve-then-reason agent maps DoD-unique cost-accounting and business-system requirements to their nearest commercial equivalents, streaming its rationale live so reviewers can follow the reasoning, not just the verdict.
- CAS & DFARS requirements
- GAAP, SOX 404(b), COSO equivalents
- Live streamed agent rationale
- Seeded, editable knowledge base
Supplier reliance assessment
For a given supplier, CLEAR derives a reliance map across every requirement — rely, retain, targeted test, or incremental — with reviewer overrides captured alongside the agent's proposal.
- Per-requirement reliance verdicts
- Reviewer override support
- Supplier-by-supplier profiles
- Nothing final without human sign-off
Reliance-dividend analytics
Reliance dispositions convert into audit-hour and dollar savings using editable assumption drivers — audit hours, DCAA rate, contractor absorption — so every figure is auditable, not asserted. Rolls up into a portfolio view across suppliers and program concentration.
- Editable assumption drivers
- Per-supplier dividend estimate
- Portfolio roll-up across suppliers
- Public-budget program exposure
Reviewer audit trail
Every agent-proposed verdict and reviewer determination is logged to an audit trail, and the deployment itself — commercial cloud, GovCloud, or fully air-gapped enclave — is a configuration choice, not a rewrite.
- Full reviewer determination log
- Swappable LLM & retrieval backends
- Air-gapped / GovCloud deployable
- CPA / legal / DCAA review required
Proposal Intelligence
BetaProposal Intelligence aggregates federal contracting data from SAM.gov, DoD SBIR, USAspending, SEWP, GSA eLibrary and more, then links vendors, awards, agencies, and technologies into an explorable knowledge graph — so business-development teams surface the right opportunities and understand the competitive landscape.
How Proposal Intelligence works
Federal source aggregation
Every relevant source, in one place.
Pulls opportunities and awards from SAM.gov, DoD SBIR/DSIP, USAspending, SEWP, GSA eLibrary, GovInfo, and DVIDS — plus curated news and RSS — normalizing each source's quirks into one consistent result set.
See detailsProcurement knowledge graph
See how vendors, awards, and agencies connect.
An embedded graph ontology links vendors (keyed on UEI), awards, contract vehicles, agencies, and technologies — explorable in-app with interactive graph views and analytics, no external database required.
See detailsAcronym-aware search
Search the way the defense world writes.
Vector search combines with defense-acronym and term expansion, so a query like C-UAS also matches “counter unmanned aircraft system” across federal records, indexed news, and live web results.
See detailsOpportunity & RFI analysis
From raw listing to a working shortlist.
Extracts the fields that matter — keyword, company, agency, timeframe, status, priority — analyzes RFIs, and exports a curated shortlist or chat transcript for the capture team.
See detailsFederal source aggregation
Pulls opportunities and awards from SAM.gov, DoD SBIR/DSIP, USAspending, SEWP, GSA eLibrary, GovInfo, and DVIDS — plus curated news and RSS — normalizing each source's quirks into one consistent result set.
- 8+ federal data sources
- Awards, notices & IDV crawl
- Curated news + live web
- Federal-first result ranking
Procurement knowledge graph
An embedded graph ontology links vendors (keyed on UEI), awards, contract vehicles, agencies, and technologies — explorable in-app with interactive graph views and analytics, no external database required.
- Vendor · Award · Agency · Technology
- Interactive graph exploration
- Concentration & relationship analytics
- Optional Neo4j / NeoDash sync
Acronym-aware search
Vector search combines with defense-acronym and term expansion, so a query like C-UAS also matches “counter unmanned aircraft system” across federal records, indexed news, and live web results.
- Semantic vector retrieval
- Defense-acronym expansion
- Relevance filtering
- Federal, news & web scopes
Opportunity & RFI analysis
Extracts the fields that matter — keyword, company, agency, timeframe, status, priority — analyzes RFIs, and exports a curated shortlist or chat transcript for the capture team.
- Heuristic field extraction
- RFI analysis
- Curated shortlist export
- Chat / transcript export
TRACE
PrototypeTRACE takes the aspirational language a SOW or RFP is written in — “operate autonomously in a GPS-denied environment” — and crosswalks it into verifiable requirements a contracting officer can hold a vendor to and a test range can score, against authoritative frameworks like DoDD 3000.09, JCIDS, and the TEMP.
How TRACE works
Testability crosswalk
Decompose aspirational language into scored clauses.
Breaks an aspirational requirement into concrete clauses and scores each on the Testability Index — how many of {trigger, threshold/metric, verification method, authoritative source} are present. “Operate autonomously” scores 0/4; the decomposed clause scores 4/4.
See detailsVerification & RVTM
From scored clauses to a draft verification matrix.
Rolls the decomposition into a per-capability scorecard and drafts a Requirements Verification Traceability Matrix — mapping each testable requirement to a verification method a test range can actually execute.
See detailsProgram risk exposure
What untestable requirements cost the program.
Turns the gap of untestable requirements into program risk — late-vs-early fix-cost delta plus test & evaluation churn — with editable cost drivers exposed in the UI so the number is auditable, not asserted.
See detailsFramework grounding
Anchored to the authoritative references.
Decomposition and verification are grounded in the frameworks that govern autonomous-systems acquisition — DoDD 3000.09, JCIDS, the TEMP, and MOSA/UMAA — with every judgment traceable to its source.
See detailsTestability crosswalk
Breaks an aspirational requirement into concrete clauses and scores each on the Testability Index — how many of {trigger, threshold/metric, verification method, authoritative source} are present. “Operate autonomously” scores 0/4; the decomposed clause scores 4/4.
- Testability Index 0–4 per clause
- Trigger · threshold · method · source
- Streamed decomposition rationale
- Surfaces the unwritten behavioral gaps
Verification & RVTM
Rolls the decomposition into a per-capability scorecard and drafts a Requirements Verification Traceability Matrix — mapping each testable requirement to a verification method a test range can actually execute.
- Per-capability decomposition scorecard
- Draft RVTM
- Verification-method assignment
- Testable / partial / aspirational bands
Program risk exposure
Turns the gap of untestable requirements into program risk — late-vs-early fix-cost delta plus test & evaluation churn — with editable cost drivers exposed in the UI so the number is auditable, not asserted.
- Untestable-req gap × fix-cost delta
- T&E churn exposure
- Editable, auditable cost drivers
- Per-program rollup
Framework grounding
Decomposition and verification are grounded in the frameworks that govern autonomous-systems acquisition — DoDD 3000.09, JCIDS, the TEMP, and MOSA/UMAA — with every judgment traceable to its source.
- DoDD 3000.09 · JCIDS · TEMP · MOSA/UMAA
- Source-cited judgments
- Air-gappable retrieval
- Reviewer audit trail
PARE
PrototypePARE (Procurement Authority & Regulatory Equivalence) crosswalks each FAR/DFARS clause in a solicitation to the authority behind it and proposes a disposition — statutory (retain), protective (retain), or discretionary (deviate). It turns the Revolutionary FAR Overhaul (OMB M-25-26) into a clause-by-clause deviation review, with the compliance burden each deviation frees priced in dollars and every call captured for a contracting officer's sign-off.
How PARE works
Clause crosswalk
Every clause, mapped to the authority behind it.
A retrieve-then-reason agent maps each FAR/DFARS clause to its governing authority and proposes a disposition — statutory, protective, or discretionary — streaming its rationale so a contracting officer follows the reasoning, not just the verdict.
See detailsSolicitation review
For one solicitation, what can actually come out?
Applies the crosswalk to a solicitation's clause set — which clauses stay because statute or an essential protection requires them, and which prescriptive clauses can be deviated away — as a draft class-deviation input.
See detailsBurden-reduction analytics
Turn deviated clauses into dollars, transparently.
Deviating discretionary clauses removes compliance hours borne by every offeror; PARE prices that with editable drivers — labor rate, offeror count, government administration multiplier — so the figure is auditable, not asserted, and rolls up across solicitations.
See detailsDetermination audit trail
Every deviation call, traced and defensible.
Every agent-proposed disposition and contracting-officer determination is logged to an audit trail — the documented basis a class-deviation package rests on — and the deployment (commercial cloud, GovCloud, or air-gapped) is a configuration choice, not a rewrite.
See detailsClause crosswalk
A retrieve-then-reason agent maps each FAR/DFARS clause to its governing authority and proposes a disposition — statutory, protective, or discretionary — streaming its rationale so a contracting officer follows the reasoning, not just the verdict.
- FAR & DFARS clause library
- U.S. Code / statute anchoring
- Statutory · protective · discretionary
- Live streamed agent rationale
Solicitation review
Applies the crosswalk to a solicitation's clause set — which clauses stay because statute or an essential protection requires them, and which prescriptive clauses can be deviated away — as a draft class-deviation input.
- Per-solicitation clause review
- Retain vs. deviate action
- Draft class-deviation memo
- Reviewer override support
Burden-reduction analytics
Deviating discretionary clauses removes compliance hours borne by every offeror; PARE prices that with editable drivers — labor rate, offeror count, government administration multiplier — so the figure is auditable, not asserted, and rolls up across solicitations.
- Compliance hours removed
- Contractor + government cost avoided
- Editable, auditable drivers
- Portfolio roll-up across solicitations
Determination audit trail
Every agent-proposed disposition and contracting-officer determination is logged to an audit trail — the documented basis a class-deviation package rests on — and the deployment (commercial cloud, GovCloud, or air-gapped) is a configuration choice, not a rewrite.
- CO determination log
- Swappable LLM & retrieval backends
- Air-gapped / GovCloud deployable
- CO / counsel review required
CEDAR
PrototypeCEDAR (Commercial Equivalence & Defense-unique Assessment of Requirements) crosswalks each requirement in a draft specification to the commercial market that could satisfy it — and to the authority that would foreclose it. Built for the commercial-first default in the Secretary of War's Warfighting Acquisition System memo and the Secretary of the Army's software acquisition memo: commercial, commercial-configured, or genuinely defense-unique, MIL-STD-961E specification type by DoW component, with the cost and schedule the defense-unique path locks in priced against real FPDS commercial-item data.
How CEDAR works
Requirement crosswalk
Every requirement, checked against the commercial market first.
A retrieve-then-reason agent proposes commercial, adaptable, or defense-unique for each specification requirement, with the governing FAR/DFARS/MIL-STD/U.S. Code citation streamed alongside its rationale — the acquisition user confirms, overrides, and records the basis.
See detailsMIL-STD-961E specification types
System, item, software, material, process — not 'hardware.'
Specifications filter by the actual DoW taxonomy — MIL-STD-961E's five program-unique specification types — and by component, so a reviewer can scope straight to their own service's software or hardware specifications.
See detailsBurden-hours analytics
The engineering hours a defense-unique path locks in, priced.
Every commercial-or-adaptable call frees non-recurring engineering hours and schedule; CEDAR prices both from editable drivers — engineering rate, NRE per requirement, schedule-delay cost — observed on the drafts in scope, then extrapolated across a year of drafting.
See detailsReal commercial-item data
What CEDAR says could be commercial, next to what FPDS says was.
USASpending's FPDS-derived commercial-item field, sampled across each component's largest FY2025 awards, sits beside CEDAR's own crosswalk — so the gap between what the memo asks for and what a component's contracting record shows is measured, not asserted.
See detailsRequirement crosswalk
A retrieve-then-reason agent proposes commercial, adaptable, or defense-unique for each specification requirement, with the governing FAR/DFARS/MIL-STD/U.S. Code citation streamed alongside its rationale — the acquisition user confirms, overrides, and records the basis.
- FAR / DFARS / MIL-STD / U.S. Code anchoring
- Commercial · adaptable · unique
- Live streamed agent rationale
- Reviewer determination & basis
MIL-STD-961E specification types
Specifications filter by the actual DoW taxonomy — MIL-STD-961E's five program-unique specification types — and by component, so a reviewer can scope straight to their own service's software or hardware specifications.
- System · item · software · material · process
- Filter by DoW component
- Scoped crosswalk ledger
- No specification hidden by a filter
Burden-hours analytics
Every commercial-or-adaptable call frees non-recurring engineering hours and schedule; CEDAR prices both from editable drivers — engineering rate, NRE per requirement, schedule-delay cost — observed on the drafts in scope, then extrapolated across a year of drafting.
- NRE hours & schedule-delay pricing
- Editable, auditable drivers
- Observed vs. extrapolated, kept separate
- Component rollup, removable vs. locked
Real commercial-item data
USASpending's FPDS-derived commercial-item field, sampled across each component's largest FY2025 awards, sits beside CEDAR's own crosswalk — so the gap between what the memo asks for and what a component's contracting record shows is measured, not asserted.
- USASpending / FPDS keyless client
- Commercial-item share by component
- Real obligations & contract-action volume
- Every figure stamped real-with-source or estimated