# Vortic

> Multi-agent AI underwriting platform and **system of action** for MGAs, Lloyd's coverholders, wholesale brokers, carriers, reinsurers, adjusters, and compliance teams.

Vortic turns broker submissions into bind-ready decisions fast: nine specialist agents run in parallel (parse, risk, flood, pricing, compliance, treaty, portfolio, memo synthesis), grounded on a data fabric (flood zones, firmographics, market signals), with Command Center as the surface where humans bind, decline, refer, delegate, and coordinate. Designed for commercial lines and delegated-authority programs that care about SLA speed **and** regulator-grade traceability.

## Agentic underwriting

Vortic uses **multi-agent orchestration** (not a single chat window): triage, parallel specialists, synthesis into an audit-ready memo, traceable runs, per-agent credit metering, and human gates before bind—suited to Lloyd's markets, MGAs, E&S, and carrier oversight.

## What Vortic does

- Parses broker PDFs and submissions, extracting insured, TIV, construction, occupancy, loss history with confidence scores.
- Runs eight specialist agents in parallel (parse, risk, flood, pricing, compliance, treaty, portfolio, memo synthesis).
- Generates a structured underwriting memo with sources cited (EA Flood, FEMA NFIP, Verisk ISO, D&B, Lloyd's Market Data, OS Postcodes).
- Surfaces decisions waiting for the underwriter via Command Center, including SLA breaches, concentration alerts, and pending decisions.
- Coordinates downstream — webhooks to brokers, treaty utilisation to reinsurers, audit trail to compliance, bound-risk references to adjusters.

## Pricing

- Three tiers: Pay-as-you-go (from $200 minimum top-up at $0.10/credit), Heavy User (from $800 at $0.08/credit, 20% volume discount), Enterprise (custom).
- $5 free credit on signup to evaluate. Credits never expire.
- Per-agent metering: parse=1, specialists=2, memo=5, full pipeline=18 credits ≈ $1.40-$1.80 per run depending on tier.
- Enterprise: SSO with SCIM, on-prem / VPC, bring-your-own-LLM, custom data fabric integrations, 1-hour SLA.

See: https://www.getvortic.com/pricing

## Architecture

- Next.js 16 App Router, Tailwind v4, TypeScript.
- Supabase (Postgres + RLS) for persistence, auth, storage.
- LLM gateway: OpenRouter, routing across a diversified stack of open-weight large language models per agent. Vendors are deliberately diversified so a single rate limit or outage does not tank the pipeline.
- Streaming via Server-Sent Events.
- Audit trail: every LLM call logged in agent_traces.
- Deploy: Vercel. Cron: daily overnight briefing at 6am.

## Who Vortic serves

- **Underwriters** — primary user; makes the bind/decline call.
- **MGAs and coverholders** — the deployment unit; binds on behalf of carriers.
- **Brokers** — submit risks, want fast quote-back.
- **Carriers** — capacity providers, want visibility into MGA decisions.
- **Reinsurers** — see treaty utilisation and concentration in real time.
- **Compliance teams** — FCA, NAIC, state regulator audit readiness.
- **Adjusters** — receive bound-risk references for claim assessment.

## Key URLs

- Home: https://www.getvortic.com/
- Platform (features): https://www.getvortic.com/platform
- Solutions (use cases per persona): https://www.getvortic.com/solutions
- Pricing: https://www.getvortic.com/pricing
- Blog: https://www.getvortic.com/blog
- Request access: https://www.getvortic.com/signup
- Sign in: https://www.getvortic.com/login

## Recent writing

- [Best AI underwriting tools compared (2026 buyer guide)](https://www.getvortic.com/blog/best-ai-underwriting-tools-compared-2026) — How to compare underwriting AI vendors: single-chat copilots, workflow automation, document AI, and agentic platforms—with evaluation criteria that map to bind speed, audit readiness, and total cost.
- [AI in insurance in 2026: practical trends teams actually adopt](https://www.getvortic.com/blog/ai-in-insurance-2026-practical-trends) — From copilot pilots to production agent graphs: where carriers, MGAs, and brokers invest in AI—and the adoption patterns that correlate with measurable bind and loss-ratio outcomes.
- [Agentic orchestration in regulated workflows (beyond chatbots)](https://www.getvortic.com/blog/agentic-orchestration-beyond-chatbots) — Design patterns for routing, parallelism, synthesis, and human gates—why orchestration layers determine whether AI survives compliance review in insurance and banking.
- [Underwriting software compared: spreadsheets, copilots, and agentic stacks](https://www.getvortic.com/blog/underwriting-software-comparison-spreadsheet-to-agents) — A plain-English comparison matrix for underwriting teams evaluating software—from Excel workflows to AI copilots and specialist agent pipelines—with Total Cost of Insight framing.
- [AI in underwriting: a production checklist for risk and compliance teams](https://www.getvortic.com/blog/ai-in-underwriting-production-checklist) — From pilot to portfolio-scale AI underwriting: documentation, evaluation, access control, and monitoring checkpoints that help legal, risk, and IT say yes without slowing bind velocity.
- [Specialist agents vs one big LLM: what works for underwriting?](https://www.getvortic.com/blog/specialist-agents-vs-monolithic-llm-underwriting) — An engineering-minded comparison of single-model prompts versus specialist agent graphs—for latency, quality, cost governance, and auditability on commercial submissions.
- [Agentic underwriting: why the future is multi-agent, not monolithic chat](https://www.getvortic.com/blog/agentic-underwriting-multi-agent-orchestration) — Single “do everything” copilots stall on complex submissions. The next wave is specialist agents, parallel execution, and orchestrated journeys — with humans at the bind line.
- [Programmable underwriting journeys (and why simulation beats slide decks)](https://www.getvortic.com/blog/programmable-underwriting-journeys-and-simulation) — Product and underwriting leads will soon design agent graphs the way they design rating rules: versioned, testable, and bound to appetite. Here is what that stack looks like.
- [Transparent agents: chain-of-thought, audit trails, and the trust layer](https://www.getvortic.com/blog/transparent-agents-chain-of-thought-and-trust) — Regulators and reinsurers do not care that you used a frontier model. They care whether you can reconstruct the decision. Agentic UX is becoming a compliance surface.
- [Toward an insurance operating system: agents, data fabric, and the bind loop](https://www.getvortic.com/blog/insurance-operating-system-agents-data-fabric) — Agents without data are toys. Data without routing is a warehouse. The winning stack connects external peril and company signals into a single action layer — fast enough for broker SLAs.
- [What is a “system of action” for underwriters?](https://www.getvortic.com/blog/what-is-a-system-of-action) — A system of record stores what happened. A system of intelligence predicts what could happen. A system of action is where decisions get made and routed.
- [Why we run on free OpenRouter models (and what to watch out for)](https://www.getvortic.com/blog/why-we-run-on-free-openrouter-models) — Llama 3.3 70B, DeepSeek V3, Qwen 2.5, Gemini 2.0 — the contrarian case for free, and the engineering you need around them.
- [The four-hour underwriter (and how a system of action gives them their day back)](https://www.getvortic.com/blog/the-four-hour-underwriter) — A composite case study based on our design partner research: where a typical underwriter spends 4 hours per submission, and which 3 hours and 27 minutes Vortic eliminates.

## Status

Early access. Free $5 credits on signup. Built for binding authorities, MGAs, Lloyd's syndicates, and US E&S coverholders.

## Contact

- Sales / pilots / design partnership: https://www.getvortic.com/contact (form) or hello@getvortic.com
- Press & analysts: press@getvortic.com
- Security disclosures: security@getvortic.com
- Early access signup (invite-required): https://www.getvortic.com/signup