A PE-backed dental group with 50+ practices across multiple states. Hundreds of clinical staff. Five different practice management systems inherited from acquisitions. When their sponsor started pushing for aggressive growth, the ops team had a problem: their back-office couldn't scale.
"Every acquisition means more systems that don't talk to each other," their COO told us. "We're already drowning in manual work. Adding more locations just means hiring more people to do the same broken processes."
Paper cuts adding up to real money
The individual tasks weren't complicated. Call the payer, wait on hold, verify coverage. Open an EOB PDF, match it to a claim, check if the payment was right. Read through a compliance binder, flag what's missing. But multiply each task by 50+ locations and thousands of patients per month, and you've got full-time staff doing nothing but administrative busywork.
The EOB problem was the worst. Two full-time billers spent their days opening PDF after PDF, squinting at explanation of benefits documents from different payers—each formatted differently—and manually matching them to claims in whatever practice management system that location used. Underpayments slipped through. Denials sat in inboxes for weeks. Their A/R aging was creeping up, and nobody had time to figure out why.
Three systems, one integration layer
We didn't replace their existing tools. The locations were used to their PMS systems, and ripping those out would've caused a revolt. Instead, we built AI that sits on top—reading what their staff would've read, doing the matching and flagging automatically.
Automated Eligibility Verification
Real-time insurance checks via payer EDI connections. When a patient gets scheduled, we verify coverage, pull deductible status, and flag anything weird—before they walk in the door. The front desk stopped calling payers entirely.
EOB Parsing and Payment Matching
AI reads every EOB that comes in—paper scans, PDF attachments, payer portal exports. It extracts line items, matches them to claims across all five PMS systems, and flags underpayments. Last month it caught $23K in short-pays that would've slipped through.
Compliance Document Analysis
150-page OSHA binders, credentialing files, policy manuals. The AI reads them, extracts required elements, flags gaps, and generates checklists. What took a compliance officer 4-8 hours now takes 15 minutes of review.
Staff doing different work, not more work
The two EOB processors? They're still employed. One handles exception cases and appeals—the stuff that actually requires judgment. The other moved to patient collections, where her knowledge of payer behavior turned out to be exactly what they needed. The front desk staff stopped dreading insurance calls. The compliance officer actually has time for proactive audits instead of scrambling before inspections.
- —15 minutes per patient for eligibility calls
- —2 FTEs manually processing EOB documents
- —4-8 hours per compliance binder review
- —Underpayments slipping through unnoticed
- —A/R aging creeping up every quarter
- 2 minutes average for automated eligibility
- EOBs processed overnight, exceptions flagged by morning
- 15 minutes per compliance review with AI pre-analysis
- $23K/month in recovered underpayments
- A/R days dropped from 38 to 29
6 weeks to production
Discovery & Design
Shadowed staff at three locations, mapped actual workflows (not the documented ones), identified the real bottlenecks.
Build & Test
Built the eligibility and EOB systems, tested against 6 months of historical data. Caught issues before they went live.
Pilot Deployment
Rolled out to 5 locations in Phoenix. Fixed edge cases, trained staff, got buy-in from skeptics.
Full Rollout
Extended to all locations. Set up monitoring dashboards. Handed off to their internal IT.
The group is on track to add 15+ locations this year. Their COO isn't worried about back-office capacity anymore. The systems scale—adding a new location is mostly just plugging in another PMS connection and letting the AI figure out the format.
The pattern here isn't unique to dental. Any multi-site healthcare operation—urgent care chains, veterinary networks, specialty practices—hits the same wall. Manual processes that work at 10 locations break at 50. The fix isn't hiring more people to do the same broken processes. It's automating the parts that shouldn't require human judgment in the first place.