Cedar · 2023
Disputes, and the phone line that takes payments
Two systems in one year: a dispute-handling pipeline that cut resolution time about 40%, and an IVR migration that lifted payments by phone 25%.
Sanitized- −40%
- dispute resolution time
- +25%
- payment-by-phone completions
- −45%
- client onboarding time
The situation
A patient disputing a medical bill is already having a bad week, and the process for handling that dispute was slow and inconsistent enough to make it worse. Accuracy and speed were both problems, and they traded against each other.
Separately, the phone system that let patients pay a bill by calling was on infrastructure that made every new client a bespoke setup.
The constraints
- Healthcare billing disputes carry real regulatory weight. "Move fast" is bounded by what you are permitted to decide automatically.
- The phone line was live and taking money. A migration had to be reversible per-client, not all-or-nothing.
The call
Give disputes an explicit pipeline with defined states and ownership at each step, instead of a queue that people worked from judgement. Consistency was the prerequisite for speed — the accuracy gain and the time gain came from the same change.
Migrate the IVR to Twilio and make provisioning a new client's phone experience configuration rather than engineering work.
What happened
Dispute accuracy improved and median resolution time fell by roughly 40%.
Payment-by-phone completions rose 25%, and onboarding a new client onto the phone experience got about 45% faster.
What I'd do differently
The two projects shared more than they appeared to — both were really about turning judgement calls into explicit configuration. I treated them as separate pieces of work for longer than I should have, and some of the dispute pipeline's state handling could have been the same machinery.