Recovering seven figures of revenue that had quietly aged off the books.

The problem

Days in A/R had drifted from 38 to 61 over four quarters. Denial rates looked acceptable in aggregate but no one had visibility into denials by payer, by service line, by provider — the slices that actually move money. Existing reporting answered yesterday's questions.

Approach

• Pulled raw 837/835 transactions, deidentified, into an analysis environment under BAA.
• Built a denials taxonomy mapped to root causes: documentation, eligibility, prior auth, coding, payer behavior.
• Identified four dominant patterns concentrated in two payer contracts and one service line.
• Delivered a remediation playbook that the existing RCM team executed against — no vendor switch required.

The AI tooling

• Pattern discovery across 14 months of denial data, surfacing combinations that human review had missed.
• Payer-behavior modeling: which contract terms were being applied inconsistently, and where leverage existed.
• Continuous monitoring dashboard that the operations team uses weekly without our involvement.

Outcome

Within ninety days, the team recovered $1.4M in previously denied claims and reduced go-forward denial rate by roughly 38%. Days in A/R returned to mid-40s.