You've tried to unclog the line before. Here's why it didn't stick.

"We have an FAQ page"

Written at the website launch, in the practice's words rather than patients', and untouched since — which means it answers questions nobody asks in language nobody searches, with insurance facts from two networks ago. The measured tell: your staff doesn't refer callers to it. When the people who know the answers won't vouch for the page, the page is furniture.

"We added a phone tree option"

"Press 3 for insurance questions" routes the clog; it doesn't shrink it. The same staff answer the same questions with an extra menu in front — and the new patient pressing 1 still waits behind the volume. Trees organize traffic. Only published answers remove it.

"We tell them to check the website"

Mid-call deflection to a page that's outdated (see above) produces a callback plus distrust — the worst of both. Deflection only works when the destination is current, complete, and in the caller's own words; otherwise "check the website" reads as "go away," which patients correctly resent.

"We looked at a chatbot"

The right instinct — self-serve answers — aimed at the wrong machine. An improvising bot that guesses about plan coverage creates liability with a typing indicator. The working version is duller and better: published, verified answers served verbatim, whether on a page, by text, or through whatever widget. Verification first; interface second.

"Front desk just handles it — that's their job"

They do, brilliantly, forty times a week — and that fluency is exactly the asset going unfiled. In our measured work, the practices' best knowledge lived in heads and vanished with turnover, vacations, and lunch breaks. "It's their job" prices skilled humans at the cost of a webpage, then wonders why the line is always busy.

What measuring changes

Each attempt aims once. The measurement counts what's actually asked, times what answering costs, and shows what it crowds out — before you spend again. That's the whole service.