# Sure Underwriting AI | Rate, quote, bind from the command line # Rate, quote, bind from the command line One AI-native CLI over your product configuration, rating tables, and policy lifecycle rules. Price quotes against filed rates, bind policies in under a second, and hand any step to an agent without losing the audit trail. [Book a walkthrough](/company/contact) [Read the quickstart](#quickstart) yourbrand.com/underwritingDraft Product definition Renters Insurance HO-4 · standard tenant policy v3.4.1v3.4.2 What's changing * NJNew JerseyDraftLive * MAMassachusettsDraftLive * ILIllinoisDraftLive Publish Under the hood ## What you get with Underwriting AI A single CLI that knows your products, your raters, your lifecycle rules, and your audit needs. Reproducible quotes. Logged binds. Reasoned decisions. #### Filed-table aware Your rate tables, lookup files, and expression formulas live inside rater versions. Publish a rater, pin it to a product, and the CLI will refuse to price against drafts that were never released. #### Sub-second policy issuance Quote, rate, issue, and write to the policy ledger in one call. The same path your product app uses, minus the browser. Payment tokenization and document generation run through the same transaction. #### Audit trail regulators can read Raters are immutable once released. Policies pin the exact revision that priced them. Export a case and you get the full factor waterfall, the rater ID, and the rule set that bound it. #### Agent-callable The CLI calls the same REST API your app uses, just from a terminal. Claude, Cursor, or any MCP-aware agent can quote and bind with the same profile, the same permissions, and the same structured errors an engineer sees. #### Scoped authority Every API call is tied to a specific partner and environment with a scoped token. If a call falls outside its bounds, it fails with a clear, structured error instead of writing bad data. #### Replay and reprice Replay a historical policy against a new rater version. Diff the breakdown lines. See the dollar impact of a rate change before you release it to a state. Quickstart ## From install to first priced quote,in six lines in six lines Quickstart · Underwriting CLI access is available to our partners. [Talk with us to learn more](/contact). Three surfaces ## Three ways to run the same underwriting pipeline Same underwriting pipeline, surfaced three ways. Pick the one your team wants to hold, and the same guardrails follow. For engineers #### CLI and API The commands you would run at a terminal, callable from your product, your CI, or your agent stack. REST under the hood, typed errors, profile-scoped auth. For product #### Embed SDK React, iOS, Android, and vanilla web SDKs that wrap the quote and bind flow. Drop in a hosted widget or wire up headless hooks and own the UI. For operators #### Admin console A web console for program managers, analysts, and compliance. Inspect any quote, replay it against a new rater, and pull audit packets without touching the terminal. Guardrails ## What the platform enforces,so you don't have to so you don't have to Every capability is bounded by filed, versioned rules. Agents can move fast because the edges are hard. 01 ##### No unreleased rates Pricing runs only against released rater versions. Drafts are visible in the filing workspace, not in production quote paths. 02 ##### Profile-scoped authority Every CLI call resolves to a profile, a partner, an environment, and a token. Cross-partner writes are rejected before they hit the API. 03 ##### Immutable revisions Raters, products, and policy subsystems are versioned. Published revisions cannot be edited in place. Changes ship as new revisions with a trail. 04 ##### Validate before publish Every product goes through rigorous automated testing before it can go live. Publish is blocked until validation passes. 05 ##### Lifecycle rules are filed Cancellation windows, reinstatement periods, effective-date offsets, and policy numbering are all configured per state and per product. 06 ##### Replay on demand Any past policy can be re-run through its pinned rater or against a new one. Breakdown lines diff cleanly, so rate changes get reviewed before release. The specs ## What it measures, in numbers Quote latency p50 62ms · p99 240ms Bind latency p50 214ms · p99 820ms Locales supported Global Lines supported All P&C, A&H, E&S, commercial, and specialty lines. Including: HO · DP · Renters · Auto · BOP · Disaster · Term life · AD&D · Home warranty Version control Raters, products, and PSVs pinned per policy revision SLA 99.97% uptime Audit format SOC 2 Type II · immutable revisions · JSON + PDF export Agent runtime Any MCP-capable LLM (Claude, GPT, Gemini) with profile-scoped tokens 214ms p50 quote-to-bind latency, measured across the full path through released raters 7LOBs Lines of business live on the platform today, from auto to life 100% Of bound policies replay cleanly against the rater that priced them > “Chubb's insurance offering, through Sure's technology platform, further reinforces our commitment to innovative, digital solutions and providing consumers with peace of mind.” James Walloga EVP Accident & Health, Chubb North America ![Chubb](/_next/image?url=%2Fpartners%2Fchubb.png&w=3840&q=75&dpl=dpl_8wGoXb7do2cSMZ3miuwhkikQx28R) [ Next ### Rating AI Your rate tables, served through Model Context Protocol and callable by Claude, any LLM, or any agent. Versioned per rater, namespaced per partner, and logged in a way you can search. Rating AI](/platform/rating-ai) ## See the platform up close Get a walkthrough with the team at Sure. We'll dig into the platform and how we impact your roadmap. [Book a walkthrough](/company/contact)