Every allied health and dental practice in Australia runs the same quiet ritual at close of day. The front desk tallies the HICAPS terminal, cross-checks it against the practice management system, and tries to work out why the numbers do not agree. A physiotherapy clinic in Parramatta or a dental group across Melbourne can lose the better part of an hour each evening to this, and the gaps rarely resolve on their own. It is unglamorous work, it happens after a full day of patients, and it is exactly the kind of task that gets rushed or skipped when the clinic is busy. Claude can take most of it off the desk.
Why HICAPS reconciliation eats so much time
HICAPS settles gap payments and rebates from private health funds. Medicare and DVA claims run on separate rails. The practice bank account, meanwhile, only records the deposits that actually arrive. That is three sources of truth for the same day of work, and they almost never line up on the same date. The reasons are predictable once you name them:
Settlement timing: HICAPS batches settle overnight, so a claim taken on Friday can land in Monday's bank deposit.
Partial rebates: a fund pays part of a fee and the patient covers the gap by card or cash, splitting one appointment across two systems.
Rejected and reversed claims: a claim that fails at the terminal can still sit in the day's takings until someone voids it by hand.
Medicare and DVA overlap: bulk-billed items and health fund items for the same visit appear in different reports.
Multi-provider clinics: several practitioners billing under one terminal makes it harder to see whose claim did not clear.
A mid-sized clinic putting $45,000 a week through HICAPS can carry more than $1,200 in unreconciled claims at any point. That is money the practice is owed but has not yet matched to a deposit. Across a year it adds up to real cash flow sitting in limbo, and it is the kind of slow leak that no one has time to trace. The cost is not only the unmatched dollars. It is the senior staff hours spent squinting at three screens, and the write-offs that happen when a mismatch is simply too old to bother chasing.
What Claude actually does with the reports
You hand Claude three files: the HICAPS settlement report, the practice management day sheet, and the bank feed. It reads all three and does the matching a human would do, only faster and without missing a line. The output is not another spreadsheet to check. It is a short list of the handful of items that genuinely need a person to decide.
Matches each HICAPS settlement line to the right appointment and the matching bank deposit.
Surfaces only the exceptions, so the front desk reads a five-line list instead of a full ledger.
Drafts the query email to a health fund, or the note for the practice manager, ready to send.
Keeps a running record so the same mismatch is never chased twice.
The practical effect is that a task which used to swallow an evening becomes a two-minute review. The staff member confirms the exceptions Claude flagged, and moves on to closing up. Nothing about the clinic's existing tools has to change. Claude sits alongside the practice management system and the terminal reports you already produce, doing the tedious cross-checking that software vendors never quite built for you.
A quick worked example
Say Tuesday's HICAPS report shows 38 transactions, but the day sheet lists 40 appointments with a fund component. Claude lines them up, finds that two claims were reversed at the terminal because the patient's cover had lapsed, and notes that one Friday claim has not yet settled into the bank. Instead of an hour of detective work, the front desk gets three plain sentences and a drafted note to the two affected patients.
Answering HICAPS queries without the phone queue
A large share of claiming admin is not reconciliation at all. It is working out what a terminal message means. Reception staff spend time on hold trying to understand a rejection code, or a rebate that came in lower than expected. Claude can read the message codes and the fund's rebate rules and explain, in plain terms, what a rejection means and what to do next. It will not replace the official support line for account changes, but it removes the guesswork before anyone picks up the phone, and it means a new receptionist is not stuck waiting for the practice manager to be free.
Keeping patient and claim data safe
Health data is sensitive, and claiming records sit squarely under the Privacy Act and the Australian Privacy Principles. A Claude-first setup can be arranged so claim data stays inside your own environment and is not used to train any model. We do not pretend that AI removes your obligations. We do design the workflow so the sensitive parts are handled the way an APRA-regulated fund or a careful practice owner would expect, with clear boundaries on what leaves the building and what does not. If you want the detail on what we do and do not guarantee, we are happy to put it in writing before anything touches a real patient record.
Most practices do not need a new claiming system. They need the admin between the systems they already run to stop taking an hour a day. If that sounds like your front desk, book a short call and we will walk through your current claiming workflow together.
