ELIAS AI — First Release
ELIAS AI is now inside FileTrac — a governed intelligence layer that searches your claims, generates your documents, analyzes your data, and guides your workflows. This is not a feature. It is the beginning of a completely different kind of claims experience.
Experience the Full ELIAS AI PlatformHow We Got Here
We didn't just add AI to FileTrac — we've been doing that since before the industry was paying attention. What ELIAS AI represents is something fundamentally different from what came before.
Then — EVO-X
The First Move
FileTrac was the first claims management system in the industry to integrate ChatGPT via EVO-X — a conversational AI layer bolted onto a live CMS. That was bold for its time. Adjusters could talk to their data. Clients were impressed. And it proved the thesis: this industry was ready for conversational intelligence.
First CMS in the industry with conversational AI
Single-model architecture — OpenAI only
Connected to FileTrac data but not governed by it
No transparency architecture — no audit trail
Proved the concept. The concept required a rebuild.
Now — ELIAS AI
The Complete Rebuild
ELIAS — Evolving Learning & Intelligent Automation Solutions — is not EVO-X improved. It is a purpose-built insurance intelligence platform designed from first principles around transparency, governance, and compliance. Multi-model by design. Human-in-the-loop by architecture. Auditable by default. This is what AI in a regulated industry actually requires.
Multi-model architecture — Claude primary, Gemini backup, Kimi K2.5 long-term
Agentic capabilities with defined authority boundaries
Every action logged, attributed, and narrated in plain language
CIS-bound — governed by the Ark Alliance Certified Integrity Standard
SOC 2 Type 2 compliant. Data isolation. No model training on your data.
ELIAS in FileTrac — First Release
The first release of ELIAS is embedded directly in your existing FileTrac platform. No new logins. No new system. The intelligence lives where your claims already live.
Natural Language Claim Search
Search your entire claims database the way you think — not by dropdown filters and date ranges, but by asking. ELIAS understands context, client names, storm events, loss types, statuses, and adjusters.
→
"Show me all open hail claims in Texas assigned to Maria from the last 90 days"
Document Generation
Generate professional documents in seconds — coverage denial letters, status updates, acknowledgment letters, loss notices — using real claim data already in FileTrac. No re-keying. No templates to maintain.
→
"Draft a coverage denial letter for claim #FT-4821 using the policy exclusion on file"
Data Visualization
Ask ELIAS for a chart and get one — instantly. Claims by status over time, reserve totals by client, cycle time trends, adjuster productivity. Your data becomes a conversation, not a report request.
→
"Show me claims received by loss type for Q1 2026 as a bar chart"
Communications Drafting
Professional correspondence — client updates, adjuster instructions, status emails — written in your voice with your claim data already populated. Review, edit, send. The writing part is handled.
→
"Write a quarterly status update for our three largest clients based on their open files"
Workflow Guidance
ELIAS knows FileTrac inside out. Ask how to configure a sub-limit, set up a fee schedule, process a batch payment, or route a claim through a specific approval workflow. Deep, contextual answers — not documentation searches.
→
"How do I configure a reserve cap for Expense vs Indemnity on a TPA contract?"
File Analysis
Upload a CSV export, a policy document, an adjuster report, or a carrier loss run. ELIAS reads it, summarizes it, answers questions about it, and — where appropriate — acts on it within FileTrac.
→
"Upload this loss run from Carrier X and compare it to what we have in FileTrac"
Human-in-the-Loop by Design
Every AI product in insurance will eventually claim to be trustworthy. ELIAS proves it. Not through marketing language, but through architecture. Every action ELIAS takes is logged, attributed, and narrated in plain language — visible to your team, your clients, and any regulator who asks. There is no black box. There is no hidden decision-making. Transparency is not a feature of ELIAS. It is the foundation it is built on.
This is what makes ELIAS deployable in a regulated industry. This is what makes it auditable. And this is what makes it different from every other AI in the market.
Full Action Logging
Every ELIAS action — query, generation, recommendation — is logged with user, timestamp, data accessed, and outcome. Tamper-evident. Queryable.
Transparency Narrative
After each session, ELIAS generates a plain-language summary of what it did — not a log dump, a readable narrative. Shareable with carriers, policyholders, and regulators.
Human Confirmation Required
ELIAS never sends, submits, or finalizes anything without explicit human confirmation. It prepares. You approve. Always.
Four Levels of AI Involvement
1
Manual
You do the work. ELIAS provides information when asked. No automation, pure intelligence-on-demand.
2
Instructed
You tell ELIAS what to do. It executes and surfaces the result. You confirm before anything is finalized.
3
Delegated
ELIAS acts on pre-approved rules. Anything outside the norm is flagged for your review before it proceeds.
4
Automated
For routine, low-risk tasks, ELIAS handles them end-to-end. Full audit logging. Every action attributed and retrievable.
You and your organization choose the level that fits. ELIAS respects the boundaries you set — enforced by architecture, not by policy. Your compliance team will want to see this.
The Ark Alliance Foundation establishes the Certified Integrity Standard — an independent, verifiable framework for how claims AI should operate. ELIAS AI is built to operate within CIS at every level. Every action is attributed. Every output is defensible. Every claim processed through ELIAS carries a transparency record that can be shown to a carrier, a policyholder, or a regulator.
This is not compliance theater. This is the only AI in the claims industry that was designed from the ground up around what the industry actually requires — and what policyholders actually deserve.
The Attribution Model — Three Elements on Every Action
Origin
Who or what initiated the action — the user, a workflow trigger, or an automated rule. Every action has a traceable starting point.
Executor
Who or what carried it out — ELIAS, a specific AI model version, or a human user. Model version is logged. No ambiguity about which intelligence was responsible.
Authority
Who is accountable for the result — the licensed adjuster, the firm, or the carrier. Accountability is established before the action is executed, not after.
FileTrac is the data authority. ELIAS AI is the intelligence layer. AdjusterIQ is where humans do their best work. Three layers, one ecosystem — and ELIAS AI is the connection between all of them.
FileTrac
The Data Authority
The single source of truth for all claim data. Built, proven, and powering $4.5M+ in annual operations. Every claim, every document, every financial transaction — structured, auditable, and accessible. FileTrac is the foundation everything else is built on.
ELIAS AI
The Intelligence Layer
ELIAS connects the data authority to the human workspace. In FileTrac, it’s the co-pilot that makes your existing data usable through conversation. In AdjusterIQ, it becomes the full product — a governed AI that follows adjusters through every step of the claims process.
AdjusterIQ
The Human Workspace
AdjusterIQ is where adjusters live — their professional identity, their verified credentials, their work history, their certifications, their marketplace for assignments. ELIAS AI is native to AdjusterIQ — not bolted on, built in.
ELIAS AI is not general-purpose AI adapted for insurance. It was designed for the specific demands of a regulated, high-stakes, documentation-intensive industry — from the model selection to the security architecture.
Built on Anthropic Claude
ELIAS is built on Claude — Anthropic’s AI — because their values align with ours. Transparency, safety, and doing things the right way. The model behind ELIAS is as important as the architecture around it.
Multi-Model Architecture
Insurance cannot afford a single point of AI failure. ELIAS runs Claude as primary with Google Gemini as backup, and Kimi K2.5 as the long-term path toward a fully internal AI system.
Your Data Stays Yours
Your claim data, policyholder information, and work product are never used to train AI models. Period. Each organization’s data is fully isolated.
Compliance
SOC 2 Type 2 — enterprise-grade security controls audited independently
Encryption
Data encrypted at rest and in transit using industry-standard protocols
Isolation
Fully isolated per organization — your data is never shared with other users
No Model Training
Conversations, documents, and claim data never used to train AI models
ELIAS AI in FileTrac is the first release — the intelligence layer inside your existing platform. The full ELIAS AI product is a separate platform purpose-built for the future of the claims industry: photo analysis, voice transcription, mobile-first field tools, full estimate assistance, and CIS-governed validation from intake to close.
ELIAS AI is live inside FileTrac now. See it search your claims, generate your documents, and guide your team — in a demo tailored to your operation.