Practice operations

"Will my insurance cover this?" - the question asked before every booking

Coverage questions arrive before almost every first booking, and they are the hardest routine question a clinic handles: the answer depends on the patient's specific policy, which the clinic cannot see, and a confident wrong answer costs the patient money and the clinic its credibility. The useful move is not to answer the question but to answer the part that is actually knowable, and to be precise about which part that is.

Separate what the clinic knows from what the insurer knows

A clinic knows which insurers it has an agreement with, which procedures it bills under which code, what it charges privately, and whether it requires pre-authorisation. It does not know the patient's deductible, their remaining annual allowance, whether their employer changed plans in January, or whether the insurer will deem a specific case medically necessary. The first list can be answered immediately and accurately; the second cannot be answered at all. Most coverage frustration comes from a clinic trying to be helpful about the second list.

Answer the knowable part without hedging it into uselessness

"We are in network with these insurers, this procedure is usually billed under this code, and the self-pay price is X" is a genuinely useful answer, and far better than "please contact your insurer". Patients asking about coverage are usually trying to decide whether to book at all, and a clear self-pay figure plus a network list lets them decide. Withholding the price because coverage is uncertain answers nothing and loses the booking.

Never let it estimate what the patient will owe

This is the boundary that must be configured before launch, not after a complaint. An assistant that produces a number - "you will probably pay about forty" - has made a financial representation on the clinic's behalf, from data it does not have. The instruction has to be explicit: state the clinic's price, state the network status, and route anything requiring the patient's plan details to a human who can verify eligibility. Ask any vendor to show you the exact wording the assistant uses at that boundary.

Pre-authorisation is where the real time is lost

For procedures needing prior approval, the expensive failure is not the question but the sequence: the patient books, arrives, and discovers approval was never requested. An assistant that knows which procedures require pre-authorisation can say so at the moment of asking, which is weeks earlier than the patient would otherwise learn it. That single intervention is usually worth more than every coverage answer it gives.

Frequently asked questions

Can an AI assistant tell a patient whether they are covered?
No, and it should be configured to refuse. Coverage depends on the individual policy, deductible and remaining allowance, none of which the clinic can see. What it can state accurately is which insurers the clinic works with, which code a procedure is billed under, and the self-pay price.
Then what is the point of letting it answer at all?
Most coverage questions are really booking decisions in disguise. A patient who learns the network list and the self-pay price has enough to decide, and gets it at two in the morning rather than after a phone queue. The clinic also stops spending reception time on a question that has a fixed answer.
What should it say when it cannot answer?
It should name what it does not know and hand over with the question intact: that the patient's specific plan details are needed, and that a named person will verify eligibility. Vague deflection - "please contact your insurer" - sends the patient away without the booking and without an answer.
Does this create a compliance exposure?
It can, which is why the boundary is a configuration decision rather than a tone preference. An estimate of patient liability is a financial representation; stating a published price is not. Keep the assistant on the second side of that line and require the vendor to show the refusal wording verbatim.
How do we know it is helping rather than deflecting?
Count how many coverage conversations end in a booking or a scheduled eligibility check, not how many were answered. An assistant that closes every coverage question with "contact your insurer" will look busy and change nothing.

Get in touch

Leave your details and we will reply. You do not need a sales call to get an answer.