Thirty days is not an aspiration; it is roughly what a product takes when the inputs exist and nobody is waiting for a vendor. The plan below assumes an MGA with a binder (or one in final draft), a wording and a tariff, and a team that can spend part of its week on it.
Week 1: inputs and decisions
- Collect the seven inputs: wording, questions, tariff, acceptance guidelines, packages, binder terms, commercial and document details.
- Decide the two things that shape everything else: packages or free sum insured, and what you are willing to bind unattended.
- Create the binder record and the product; set branding and letterhead once.
Week 2: build and prove the engine
- Fill in the product workbook: Questions, Rates and Rules. Upload, read the checks, confirm.
- Convert the tariff carefully (additive versus multiplicative, minimum premium, package prices) and rate the ten test cases on the Rating tab until they match the old tariff.
- Walk the rules with edge cases on the binder limit and every knock-out question.
Week 3: documents, people and channels
- Upload the wording, preview the quote and schedule PDFs, check the invoice and the e-mails.
- Invite the team with the right presets and switch on Require MFA.
- Set up the first broker organisations and decide which channels open at launch: broker portal first, quote page and API when the product has proven itself.
- Connect your domain if brokers should never see the platform's address.
Week 4: pilot and go-live
- Run a real submission end to end with one friendly broker: submit, refer, quote, bind, invoice, and check the policy appears in the bordereau preview.
- Fix wording, order of questions and rule thresholds from what the pilot shows; these are workbook edits, not change requests.
- Tick 'Visible in the broker portal' for everyone, send the portal link, and schedule the first bordereau.
What usually slips, and how to avoid it
- The tariff is not written down
- Get it as rows before week 2 starts. A tariff 'in the model' is the single most common cause of a slipped launch.
- Too many questions
- Every question is friction for the broker and data you must protect. Keep those that rate or decide.
- Waiting for the perfect wording
- Launch with the approved wording; upload the next version on the Channels & terms tab when legal is done. Issued policies keep theirs.
- Nobody owns the rules
- Name one underwriter who owns the Rules sheet. Referrals are a design choice, not an accident.
Related