About Zorva
Zorva is built and operated by Ashbi Design — a small
Ontario studio that ships conversion-focused websites for
service businesses and ecommerce brands. Zorva is the
studio's first healthcare product, and the first
Alberta-first AHCIP pre-submit auditor built by an
independent shop.
Get in touch
The story
Zorva started as a small internal tool inside Ashbi Design —
we noticed that the same 5-10 AHCIP billing patterns were
causing the same 5-10 denials, on every engagement we did
with Alberta clinics. We tried every off-the-shelf billing
tool on the market. None of them read the clinical note
against the AHCIP fee schedule. None of them caught the
modifier-25 problem. None of them surfaced the CMGP
underpayment.
So we built one. v12 of the Zorva prompt ships with 18 AHCIP
rules — every one of them something an Alberta biller
needs to know about every claim. The auditor runs at
F1=0.690 (precision 0.625, recall 0.769) on a 10-encounter
/ 13-gold-finding cleaned AHCIP val set. Small-n validation,
we know — we're working on the next iteration with a
larger partner-clinic sample. The full metrics are
published in how-it-works.
We are not trying to be the largest billing tool. We are
trying to be the one Alberta billers keep open in a tab all
day.
The team
Cameron Ashley — founder
Cameron has shipped 80+ projects over the last five years,
with a Top Rated badge on Upwork. He has built and
maintained websites for clients in HVAC, ecommerce,
law, and a dozen other verticals. He lives in Toronto.
Before starting Ashbi, Cameron worked as a software
engineer at a major Canadian healthcare technology
company. He has spent the last year focused almost
exclusively on the AHCIP billing problem.
Zorva v12 — the auditor
The auditor is a large language model, prompted to read
the clinical note + the billed claim + the applicable AHCIP
rules and emit a structured finding list. The model is
brought in by the customer — Zorva ships the prompt,
the rules, the audit trail, and the biller's review UI. Your
data does not train anyone else's model.