Who it is for
Rate at the edge of your channel, on the version you released
The people quoting your business are mostly not your staff. They sit one, two or three steps away from the system that holds the rates, and every step is a chance for the number to be stale, wrong or outside the authority you granted. SelectRate puts one endpoint at the center of that channel and lets you decide exactly which programs each partner can rate.
Why one page covers wholesalers and program administrators
The contracts are different. A wholesaler places business it does not control with carriers it does not own, through retail agents who work for somebody else. A program administrator runs a defined program under delegated authority and is measured on how that program performs. Those are not the same job.
The rating and distribution problem, though, is the same shape: your economics depend on people outside your organization producing correct numbers on your released version, inside authority you granted and can prove you granted. Everything below applies to both, and where it genuinely differs, the page says so.
Distance is the problem
A rate produced inside your own underwriting team is easy to trust. You can see it happen.
Two steps out, trust stops being observation and becomes assumption. The retail desk is quoting from a spreadsheet somebody built from your rate pages last year. An MGA writing under you has its own copy of your rules and its own idea of when it last updated them. A partner is quoting a class you never granted, and nobody noticed because nothing stopped them.
You usually find out in one of two places. At bind, when the number does not hold. Or at audit, when somebody asks how a rate was produced and the honest answer is that it was produced outside your system.
That is not a discipline problem. It is an architecture problem. If the rate can be copied, it will be copied, and the copy will age.
One endpoint, three kinds of user
One endpoint serves your own underwriters, the MGAs who write under you, and their retail agents, all on the version you released.
| Who | Where they work | What they get |
|---|---|---|
| Your own underwriters | Underwriter Workbench | Full rating and comparison on real submissions, with the worksheet behind every number. |
| Sub-producer MGAs | Their own API client | Programmatic rating limited to the programs you granted them. |
| Retail agents | An authorized rating channel you run | A rate from the same engine, reaching them through a channel you control rather than through a copy of your rates. |
Workbench and the Rating API run the same engine. Identical inputs, program and effective date produce an identical premium and an identical worksheet. The number an agent sees at a desk you have never visited is the number your own underwriter would have produced from the same facts.
Authority that travels with the rate
Give each downstream partner its own API client. Each API client has its own credential and is granted the specific programs it may rate. Scopes decide which operations a credential can call. Grants decide which programs. There is no wildcard at request time, and a client never sees another client's quotes or policies.
The enforcement model is deliberately blunt. Scopes gate verbs, so a credential can be allowed to quote, to read, or to read worksheets, separately. Grants gate programs. There is no wildcard at request time, so nobody can ask for everything and see what comes back. A program a partner is not entitled to returns 403 without disclosing that it exists, so a partner cannot map your book by probing for programs it was never given.
The distinction that matters here is between authority written down and authority enforced. A binding agreement describes what a partner may rate. Grants decide what a partner can rate. When those two are the same object, the quarterly exercise of checking whether anyone has strayed outside appetite mostly stops being necessary.
Every call leaves a record
A quote holds revisions, a revision holds rating runs, a run produces a result, and the execution is immutable. Amending a quote increments the revision; rating does not. Old runs are retrievable and byte identical forever, every call carries a correlation id, and an Idempotency-Key header is supported so a retry at a flaky retail desk does not become a second quote.
Worksheets are not inlined, because a run can carry seventy or more steps. Each result links to its own worksheet, and the worksheet holds every step as the engine executed it, with the governing value and where that value came from, whether that is a filed source, a carrier version or a program version.
Which version produced a given rate is not a question a partner answers. The effective date is the only temporal input the API accepts, so the version follows the policy date. Each partner rates under its own credential, and every rating run stays retrievable exactly as it was executed, with a correlation id. Ask to see how that record is surfaced for a channel your size before you design a reporting process around it.
No spreadsheets in the field, no stale rates at the agent desk
When you change a program, the new version is published with an effective date and the change is live in Workbench and on the Rating API at once. There is nothing to distribute, no file to send, and no interval during which half your channel is quoting last month's position.
A program that is not effective for the quote date returns no premium. At the edge of a channel that is the behavior you want, because the alternative is a confident number produced from the wrong version by somebody who had no way of knowing.
And when a rate has to be explained later, prior versions stay callable for audit and a quote from three years ago re-rates on the version and edition it was written under. A renewal an agent quoted two years ago is still explainable without anybody reconstructing it.
Adding a partner, adding a program
Onboarding a partner is a grant, not a project: an API client with its own credential, and the programs it may rate.
Standing up a new program is configuration. You configure it in Program Studio and test it there before it goes forward. You configure it. Not a support ticket. The panel view of that, rating one risk across several carriers at once, sits on the MGA page.
What stays yours
SelectRate rates and records every transaction on the policy term. Billing, invoicing, commissions and accounting stay in your systems. Your portal stays your portal.
Endorsement, cancellation, reinstatement and renewal are rated in SelectRate from the policy's current accepted state, on the rating versions locked at the term's effective date. Every accepted transaction is kept, unchangeable, in the policy's transaction ledger.
Rating is synchronous, with no webhooks.
Questions wholesalers and program administrators actually ask
Can a retail agent see the worksheet behind a rate?
Only if you allow it. Worksheet reading is a separate scope from quoting, so you can give a sub-producer the ability to rate without giving it the full step-by-step behind the number.
Can we stop a partner rating a program we never granted?
Yes. Each API client has its own credential and is granted the specific programs it may rate. Scopes decide which operations a credential can call. Grants decide which programs. There is no wildcard at request time, and a client never sees another client's quotes or policies. An ungranted program returns 403 without disclosing that it exists.
Do our partners need to track versions?
No. There are no version identifiers in the public contract and no way to pin a version. Partners send an effective date, and the engine resolves the rest.
Can partners send us batches or receive callbacks?
Not in this release. Rating is synchronous, with no webhooks and no asynchronous or batch rating. That is published rather than discovered later.
Bring the messiest part of your channel
The useful demo is not your cleanest program. It is the partner arrangement you would struggle to describe in a diagram. Book a call, bring a program you already distribute, and we will configure it with you while you watch.