In a broking office the work moves between the broking system, insurer portals, email and the phone. Account executives and claims staff retype what clients send into an insurer’s form or portal, then retype the insurer’s answer back into the client file. Claims, renewals, premium chasing, new client set-up, complaints and the shared inbox all run this way, and the file notes are often the only record of who said what. When the predictable steps run as fixed, tested software, reminders, status notices and record updates happen on schedule without anyone remembering them. AI is used only where something has to be read or drafted, such as a loss description or a complaint reply, and its output is checked before anything reaches a client or an insurer.
These are patterns we see in insurance broking businesses, not client case studies. Your process gets its own map in the assessment.
Claims notifications and insurer follow-up
A client rings after a storm with water through the ceiling, then sends photos, a plumber’s report and a repair quote over the next few days. The claims officer lodges the notification with the insurer, chases the claim number, forwards documents as they arrive and fields the client’s calls. AI reads the loss description and each incoming document against the claim checklist, so the officer can see what is missing without opening every attachment. The routine around it runs on a schedule, sending documents to the insurer by portal or email, requesting updates and sending status notices to the client. Decisions on the claim, and any advice to the client, stay with the claims officer.
Chasing premiums before cover is at risk
Premium invoices go out ahead of renewal, and an unpaid premium can put a client’s cover at risk. The accounts team chases each one, matches payments that arrive with the client’s own reference instead of the invoice number, and follows up premium funding arrangements. Basic reminders already exist in most accounting and broking systems, so the part worth automating is the matching and the consistency of follow-up. Payments are matched to invoices on rules, reminders go out on the cadence you set and each contact is logged. AI reads remittances, classifies client replies and drafts any non-standard follow-up for the accounts team to approve. A client in hardship or disputing the premium goes to their account executive.
Winning and setting up clients
Quote enquiries reach a broker through the website form, referral partners, phone calls and emails forwarded by existing clients. Plenty of CRM and broking software already captures web forms well, and if yours does, this is a small job. The manual work is usually in the other channels, where a referral email has to be typed in, a phone enquiry is written on a notepad, or a duplicate record is created for someone who is already a client. Each enquiry is captured into the broking system or CRM, checked against existing clients and assigned to an account executive. The executive then does the part only they can do, which is calling the prospect.
Once a new client signs the letter of appointment, someone creates them in the broking system, records the entities, contacts and insured locations, links any policies transferring from the previous broker and sets up billing. The same details often go into the accounting system and a client portal as well, typed again each time. No AI is needed, and this is fixed software from end to end. Details captured at enquiry and in the signed appointment flow into each system, with checks that required fields are complete and consistent across them. The account executive is told when the file is ready, and any gap between systems comes back to them to resolve.
Inboxes and complaints
Certificate of currency requests, mid-term changes, claim documents, insurer endorsements and renewal invitations all land in the same broking inbox, most of them under a vague subject line. Whoever is on inbox duty reads each email, finds the client and policy in the broking system and forwards it on. In the automated version AI classifies each message by type and pulls out the client name, policy number and insurer. Routing rules send it to the right account executive or claims officer, or attach it straight to the client file. A message the classifier cannot place with confidence waits for whoever is on inbox duty, and it stays in that queue until they take it.
A client unhappy with a claim decision rings their account executive, who may or may not write it up as a complaint. Your complaints policy sets response timeframes, and meeting them depends on someone logging the complaint on the day it arrives. Each complaint is captured from email, phone notes or the web form, the clock your policy sets starts, and the owner is reminded as each deadline approaches. For the reply, AI drafts a response from the file notes and correspondence for the complaints officer to edit. Nothing goes to the client until the complaints officer has approved it, and the register shows every complaint, its status and each response sent.
Renewal dates and privacy records
A broker’s year is built around renewal dates, and around the obligations in each insurer agency agreement and binder, which set reporting dates, notice periods and authority limits. Renewals usually live in the broking system. Everything else tends to sit in a spreadsheet that one person keeps up to date, taken from agreements that run to long PDFs. AI reads each agreement and extracts the dates, notice periods and obligations, and the person who used to keep the spreadsheet checks that extraction before it enters the register. From there the register runs on rules, sending reminders ahead of each date to whoever owns it and escalating when nobody acts.
From 10 December 2026, APP entities must disclose in their privacy policy the kinds of personal information used in, and decisions made by, computer programs that could significantly affect individuals. Brokers hold sensitive personal information, including claims histories, health details for some covers and financial records, spread across the broking system, email, insurer portals and shared folders. Fixed software keeps a register of which systems hold what and which programs use it, and tracks each access request through to its response date. When a client asks for their records, AI helps find them across those systems, and a staff member reviews every response before it goes. Your business keeps responsibility for its privacy policy.
Retyping between insurer portals and the broking system
A commercial policy can be keyed into an insurer’s portal for a quote, into the broking system when it binds, and into the accounting system when the invoice goes out. Endorsements and mid-term changes repeat the cycle. Each step is small, so nobody counts them, but the mismatches surface later as a wrong premium on an invoice or a certificate of currency with an old address. Most of this is fixed software, with mappings between systems, scheduled sync where the insurer allows it, reconciliation checks and a queue for records that disagree. AI maps free-text notes into the right fields, and those mappings are reviewed before they are saved. We measure the whole thing as one process before anything is built.
If one of these costs you real money every month, describe it in four answers.
Thinking about what a process actually costs you?
The assessment is small, fixed-price, and tells you the real number, whether or not you ever build.
Worth a 30-minute conversation.