You bring a piece of work your firm does over and over. We design the workflow against how that work is really done, connect it to your documents and systems, build it, test it on real work, and hand it to the people who own it. One fixed fee covers the design and the build.
A brokerage wants renewal notes prepared from the expiring schedule, the claims history and the client's updates, so the broker reviews rather than assembles. The stages below are the same for every workflow; the details are this one's.
Illustrative example
Time with the brokers who do renewals, not the process document. Which documents arrive and from where, what the broker checks first, which cases go to a senior broker and why. The controls model is written here: what the AI may read, what it may draft, and the three conditions under which a renewal is held for a person. You sign the design off before anything is built.
The broking system and the document store are connected through connectors scoped to renewals only. The workflow prepares the renewal note from the schedule, the claims history and the client's answers, cites the source for every figure, and writes the AI Workpaper as it goes. Nothing is sent to a client or an insurer by the workflow.
An evaluation set agreed with the workflow owner: closed renewals with known outcomes. The output is compared with what the brokers actually produced, the misses are fixed, and the holds are checked to fire on the cases that should stop.
The scope named them at the start: the note prepared for the agreed renewal types, sources cited, the hold conditions firing, the workpaper readable by the compliance manager. Acceptance is a comparison against that list, not a judgement call on the day.
The team is trained on the workflow and on what to check. The documentation describes how it works in the brokerage's own terms, who owns which control, and what to do when it holds a case. The engagement ends with the owner running it, not with a demonstration.
The scope says who to call, for how long, and what counts as a fix rather than new work. If the brokerage wants the workflow watched and improved beyond that, or a second workflow built, that is agreed separately.
Not a chatbot. The model does the reading, comparing and drafting; the platform around it records where each case has reached, keeps the evidence, and pauses the work where a person has to decide. Your people keep working inside the assistant they already use, and make the calls that need professional judgement.
Most insurance and financial-services work clusters around document-heavy review and preparation. The industry pages show one worked example each.
Submission packs, prior policies, claims history, missing-information detection, evidence-based recommendations.
See the example → Insurers & MGAsChronology, key issues, outstanding actions, evidence and citations across large claims files.
Insurers & MGAs → BrokersCompare current and prior year, identify changes, draft the renewal note, hold the cases that need a broker.
See the example → BrokersClassify shared-mailbox traffic, detect urgency, flag missing documents, draft responses.
Brokers → AccountingSchedules rolled forward from the ledger with the source, the check and the reviewer's sign-off on the file.
See the example → Finance teamsInvoices classified, duplicate-checked, coded from supplier history and drafted in Xero; new bank details stop for a person.
See the example →One workflow at production quality beats five demos that never ship. The first workflow is chosen on evidence from the first conversation, not from an open whiteboard.
Document repositories, CRMs and core platforms stay the systems of record. The assistant becomes the interface to them.
Who must approve what is decided in design, and those checks are tested before go-live. Governance is not a phase that follows the build.
Where the assistant itself can be the workbench, we let it. Custom screens only where they earn their place.
The technology is the easy part. An implementation sticks when the owner can run the workflow and the reviewer knows what to check.
Where AI could make the biggest difference, how that work runs today, the systems it touches and who owns it. We also work out together whether Airclerk is the right partner for it. If we are, the conversation gives us enough to quote.
One document: the outcome we are aiming for, what we deliver, what your team contributes, the fixed fee and timeline, and what acceptance looks like. Nothing starts until it is agreed on both sides.
The engagement opens with the detailed design, agreed with you before anything is built. Then we configure and connect the workflow to your documents and systems, build in the agreed review controls and audit trail, test it on real work, and prepare the people who will run it.
Handover, and the support that follows go-live, are written into the scope so you know what is covered before you sign. Ongoing operations and further workflows are a separate agreement, made when you want them rather than switched on by default.
A free introductory conversation is enough to know whether it is a first workflow and what it would cost to build.
The agreed detailed design; the workflow working on your real documents and systems; the connectors it needs; a governance and controls model for that workflow; the run record the controls model calls for; the people who own it trained and running it; and documentation and handover written for your firm. Testing, acceptance and the support that follows go-live are set out in the scope.
It is tested on real work before anyone relies on it: past periods, closed matters or de-identified files, with an evaluation set agreed with the workflow owner. Acceptance criteria are written into the scope, so acceptance is a comparison against them rather than a judgement call on the day.
The person who owns the workflow day to day, with time in the design phase and at testing; someone from risk or security to sign off the controls model and access; a sponsor; and access to the documents and systems the workflow needs, under your own platform licences.
The shape of the workflow: how many systems it touches, whether connectors exist or have to be built, how many decision points need a person, and how much of the work is judgement rather than mechanics. A document-only workflow and one that crosses several systems are different builds. The fee is fixed in NZD plus GST and quoted in writing after the first conversation; platform licences and usage are separate.
The support that follows go-live is written into the scope, so you know what is covered before you sign. Ongoing monitoring, improvement and further workflows are a separate agreement, not switched on by default.