LibreAnalyticsGet in touch

Data analytics · AI research · Automation

AI that shows its work.

Libre Analytics is a data analytics and AI research firm. We build automation for high-stakes work where confidence isn’t enough: every decision, figure and action our systems produce can be cited, sourced and traced back to how it was reached. In any industry.

See how we work Investors and partners

Why we exist

Every industry runs on the same hidden layer: documents, handoffs and decisions.

Claims, purchase orders, shipments, contracts. The systems of record are rarely where the work happens. It happens in inboxes, PDFs, spreadsheets and the people who reconcile them. That layer is where we do our research, and where automation actually pays.

The paper trailLibreA record you can check
  1. 01

    The bottleneck is never the data.

    It’s scattered across inboxes, documents and disconnected systems. The work is reading it in context, every time it changes.

  2. 02

    Domain knowledge is the moat.

    Models are everywhere. Knowing what a supplement, a short-ship or a change order actually means is not. We learn the domain before we automate it.

  3. 03

    Confidence isn’t the bar. Auditability is.

    In high-stakes work, “the model is usually right” is not an answer. Every output needs a source, a reason and a record someone can check.

Observable · Auditable

The audit trail is the product.

Most AI asks to be trusted. Ours is built to be checked. Every system we ship records what it read, how it reasoned, what it decided and what it did, so a person, a regulator or a client can follow any output back to its origin.

Cite
Every claim points to the exact message, page or line it rests on.
Source
Every figure keeps its lineage: where it came from, when, and what changed it.
Trace
Every decision keeps the reasoning steps that produced it, in order.
Audit
Every action is logged with what, when and who approved it, and can be replayed.
Trace inspectorIllustrative
Output

Supplement requested: roof decking, $14,860

Reasoning
  1. Carrier estimate omits decking replacement
  2. Engineer found rot across 22 sheets
  3. Policy covers code-required repairs
Evidence
  • Carrier estimate · line 31
  • Engineer report · p.4
  • Policy · p.12, Ordinance or Law
Actions
  • Supplement draftedSystem
  • Sent for approvalAwaiting adjuster
  • People hold the penConsequential actions wait for a person’s approval, and the approval becomes part of the record.
  • Least access, alwaysSystems read what they need and change nothing on their own. Each client’s data is walled off from every other.
  • No trail, no shipIf a system can’t show its work, it doesn’t go to production. That rule holds for our own products first.

What we do

We research the work, then build the system that does it.

Libre is industry-agnostic by design. The same method (study the workflow, model the data, automate the handoffs, prove every output) carries from one domain to the next. Each system we build makes the next one faster.

  1. 01

    Study

    We map how work actually moves: the documents, decisions and people involved, and where time disappears.

  2. 02

    Model

    We turn that paper trail into structured, connected data, with analytics that show what is really happening.

  3. 03

    Automate

    Agents and integrations take over the reading, reconciling and follow-up, inside the tools teams already use.

  4. 04

    Prove

    Every output is cited, traced and logged, then measured against the old way. If it doesn’t pay, we say so.

Data analytics

Audits, pipelines and decision analytics built on the data an operation already produces.

AI research

Document understanding, retrieval and domain reasoning, including our own model, Clarity.

Integrated automation

Systems that connect email, ERPs, CRMs and files so work flows without re-keying.

Where the pattern repeats

  • Insurance claims
  • Restoration
  • Supply chain
  • Procurement
  • Logistics
  • Construction
  • Manufacturing
  • Finance operations
  • Field services
  • Compliance

Clarity · Alpha

The reasoning layer beneath everything we build.

Clarity is Libre’s proprietary model, built for the documents, decisions and rules that complex operations run on. It researches, finds answers and builds strategy with the full context of the work behind it. Its first proving ground is property claims.

In alpha with select partnersAsk about the partner program
  • Fluent in the paper trailPolicies, estimates, letters and reports, read the way the profession reads them.
  • Reasoning you can inspectAnswers cite the page and passage they rest on, and expose the steps that got there.
  • Proven in-house firstThe agents that build, test and run Libre’s products, and the systems that run the company, run on Clarity every day.
  • One core, every domainBuilt to carry across industries: the Edge Series first, then whatever we build next.

Research · Supply chain

Supply chain automation and analytics.

Supply chains run on the same mess of email, spreadsheets and exceptions that claims do. Drawing on years of global supply-chain and analytics transformation work, we’re exploring how source-traced AI applies to planning, procurement and logistics.

Explore a research partnership
ResearchAreas we’re exploring
  • Exception triage from email and EDI
  • Supplier and PO reconciliation
  • Planning signals with their sources attached

In market · The Edge Series

Two products. One national view of how property losses happen and get paid.

LeadsEdge researches losses from the restoration and construction side, across the country. AdjustEdge manages and adjusts the claims that follow. Building both puts us inside the whole life of a loss, and what we learn there shapes every release. That is why the Edge Series moves as fast as it does.

  1. 01 · The lossA loss happens

    Fire, wind, hail or water hits a property, somewhere in the US.

    LearnsWhere and how losses happen: peril, property type, region, season.

  2. 02 · LeadsEdgeResearched

    Property, owner, event and code researched in under three minutes, every field cited.

    LearnsWhich properties, codes and conditions drive the size of a loss.

  3. 03 · LeadsEdgeScoped and priced

    Restoration scope and replacement cost worked out, with the math shown.

    LearnsHow losses are scoped and priced, region by region.

  4. 04 · AdjustEdgeClaim worked

    Every email, document and figure kept current as the claim moves.

    LearnsHow claims actually move: who acts, when, and what stalls them.

  5. 05 · AdjustEdgeAdjusted and paid

    Supplements, disputes and settlement, each figure traced to its source.

    LearnsWhat gets paid, what gets disputed, and why.

  6. 06 · Both productsEvery release gets sharper

    What we learn across the life of a loss shapes the next release of both products.

    LearnsA deepening, national understanding of property loss, from first call to final payment.

EdgeOS 2 · Fall 2026

AdjustEdge

Claims software for public adjusting firms

Every moving part of your claims, kept current from your inbox.

EdgeOS 2 reads each claim’s email and keeps its work, money, people, files and timeline up to date by itself, with the source behind everything.

  • Reads claim email from Gmail and Outlook, read-only
  • Every figure traced to the email that stated it
  • Supplements, disputes and settlement tracked to the figure
  • Private by firm: your claims are yours alone
Schedule a demo Visit adjustedge.ai
Live · Rolling out

LeadsEdge

Post-loss research for restoration and public adjusting

One address in. A researched, cited readout out.

Enter an address and a peril. LeadsEdge fuses live public data into one loss readout: the property, the owner, the event, code triggers and the math, with every field grounded to a source.

  • Eight live feeds fused per run
  • Every field cites its source URL and a trust score
  • Residential and commercial research modes
  • Built for restoration firms, GCs, roofers and public adjusters
Request access Visit leadsedge.ai

Investors and partners

A research firm that ships products.

Libre ships real products to real customers, and keeps improving them at a pace few teams match. Underneath is one shared core (document understanding, source tracing, integration and domain reasoning), so each new vertical starts further ahead than the last.

Portfolio stage: AdjustEdge is live, LeadsEdge is live and rolling out, Clarity is in alpha, supply chain is in research.
AdjustEdgeFive releases since Feb 2026
LeadsEdgeRolling out to restoration and adjusting firms
ClarityProprietary model · select partners
Supply chainAutomation and analytics

Traction · Unit economics

Real products, real customers. Profitable, and faster each time.

From first idea and prototype to a profitable product in seven months, then in four. Plenty of AI products cost more to run than they earn; ours don’t. AdjustEdge turned profitable at its second release and LeadsEdge at its first, and both keep improving every few weeks, because our research goes into features that pay back, not features that demo well.

1month
AdjustEdge: launch to profitable
At launch
LeadsEdge: profitable from its first release
7 → 4 mo
Idea to profitable product: first vs second
7releases
shipped across both products since Feb 2026

Idea + prototype → launch → profitable

AdjustEdgeProfitable at V25 releases shipped
Idea + prototypeProfitableV1V2V3V4V5 · EdgeOS 27 months, idea to profit
LeadsEdgeProfitable at V12 releases shipped
Idea + prototypeProfitableV1V2V3 · mid-Oct4 months, idea to profit

Profitable: the product’s revenue exceeds the direct cost of running it for customers (AI models, data and tooling), on our own lean, agent-driven operation. Figures available to investors on request.

779
claims managed in AdjustEdge
4,360
claim documents read in context, 137,000+ pages
227,000+
emails screened, 8,000+ filed to claims

Defensibility

What a model provider can’t ship as a feature.

Investors now ask every AI company the same question. Our answer runs up the full ladder of vertical-AI moats, and the higher rungs are where Libre is built.

  1. 01Owned workflow

    We run the daily work itself: claims kept current from the inbox, losses researched from a single address.

    Live
  2. 02Vertical integrations

    Wired into where the work already lives: Gmail and Outlook, and eight live public data feeds per research run.

    Live
  3. 03Built-in auditability

    Every output cited, traced and logged: the layer regulators are now asking insurers to govern.

    Live
  4. 04Domain depth

    Two products inside how property losses happen, are scoped, priced, adjusted and paid: an understanding no general-purpose model has.

    Live
  5. 05Embedded judgment

    Clarity turns that context into domain reasoning that compounds with every loss.

    Alpha

Moat framework: Mandal & Sinha, “Defensible Moats for Vertical AI Application Companies,” Stanford Law, June 2026

30
US jurisdictions now set AI governance expectations for insurers: 26 have adopted the NAIC AI Model Bulletin, 4 have their own rulesNAIC adoption map, Aug 31 2026
~75%
of global insurtech funding in Q1 2026 went to companies with AI at the core of the productGallagher Re, Q1 2026
$222B
in US AI and ML venture deal volume in 2025, up 105% year over yearPitchBook-NVCA, via Houlihan Lokey
  • Auditability is becoming a requirement. Thirty US jurisdictions now expect insurers to govern their AI with transparency and accountability. We build the audit trail in from the first line, not as a retrofit.
  • Pilots stall; operations don’t wait. Most AI never leaves the demo. We start inside a real workflow and stay until it runs.
  • Parallel verticals, compounding position. LeadsEdge works losses from the restoration side and AdjustEdge sees them adjusted and paid, nationwide. Building both shows us the whole life of a property loss, which is why both products have improved so quickly.
  • Lean by design, profitable without outside capital. Two products and seven releases in eight months, built and run on our own agent infrastructure. The company itself runs on software we engineered, from product management to finance. Capital would go to reach, not to keeping the lights on.
  • Model-agnostic by design. Our products run on several leading AI providers and can move work between them, so no single model vendor sets our costs or our ceiling.
  • The core compounds. Reading, tracing, integration and reasoning are shared across every product. Each new product reached profit faster than the last.
  • Domain first, then model. Products earn the workflow and the data context. Clarity turns that context into proprietary reasoning.

We’re profitable and not raising on a deadline. But we’re always glad to meet investors and partners who see what we see, and the right ones help us reach the next vertical sooner.

Start a conversation Ask for our overview

Our story

Reclaim the time that operations stole.

We started because we kept seeing the same thing: smart people in complex industries spending half their day on work a well-built system could handle in seconds. Not because they lacked talent. Because the tools weren’t built for their actual job.

“Libre” means free. Free from the operational drag that holds organizations back. Free from the false choice between speed and trust. Free to work at the highest level of your talent, because the systems behind you handle the rest.

We research. We build. And when the pattern is clear and repeatable, we productize. AdjustEdge and LeadsEdge both began that way: a real workflow studied up close, then turned into a product that pays for itself.

And Libre runs on what it builds. Product management, finance and email automation all run on systems we engineered, and our products are built, tested, shipped and supported by our own agent infrastructure on Clarity, with a person approving every consequential step. The standard we build products to is the one we run the company on: deliberately lean, fast and profitable.

  • The heroKokopelli

    A symbol of freedom. In the Libre universe, the hero who helps organizations break free from decision latency and reclaim stolen time.

  • The antagonistThe Time Thief

    Everything that consumes time inefficiently. Always late, yet in a constant state of urgency, and the reason good teams make rushed decisions.

What we hold to
  • CuriosityThe door

    We ask questions before we assume answers, and explore before we prescribe.

  • ClarityThe prism

    Complexity is easy. Clarity takes real work. If it takes a PhD to use, we built it wrong.

  • GrowthThe bloom

    When a week of work takes a minute, the rest of your time opens up.

  • ExcellenceThe atom

    We don’t ship things we aren’t proud of. If we can’t do it at the highest level, we don’t do it.

  • IntegrityThe scales

    Every capability we claim, we can demonstrate. Every number we cite, we can source.

Our visionHelp the world break free from decision latency.

In market
Insurance and restoration, across the US
Approach
Observable, auditable automation