Ambulatory surgery centers have historically trailed hospitals in revenue cycle technology. Smaller margins, leaner back offices, and high per-case coding complexity made large RCM platforms a hard sell. That gap is closing fast, and the catalyst is agentic AI — software that doesn’t just flag a code for a human to check, but reads the operative note, assigns the code, and routes the claim with limited human intervention.
The ASC coding bottleneck
Surgical coding is unusually dense: a single case can touch multiple CPT codes, modifiers, and implant or device line items, each with its own documentation requirement. ASCs run this process with a fraction of the coding staff a hospital system has, which means backlogs are common and first-pass claim accuracy varies widely by facility. That bottleneck is exactly what agentic AI vendors are now targeting.
What’s actually being deployed
Documentation-to-code pipelines
The newer platforms ingest the operative note directly — sometimes via ambient transcription at the point of care — and generate a coded claim without a coder manually abstracting every line. The pitch is fewer manual touches between the surgeon’s note and the submitted claim, which in theory shortens the time to reimbursement.
Claim-level automation beyond coding
Several ASC-focused deployments pair coding automation with eligibility checks, prior authorization routing, and payment posting, treating coding as one stage in a fully automated claims pipeline rather than an isolated task.
A concrete example: the CBFS–Milagro deployment
One of the clearer signals of this shift is Constitution Billing and Financial Services (CBFS), the billing subsidiary of the Constitution Surgery Alliance (CSA) network, signing a 2025 commercial agreement with AI vendor Milagro to deploy its autonomous coding platform across CSA’s partnered ambulatory surgery centers, integrating directly with the Surgical Information Systems EHR. CBFS described the goal as reducing manual coding touches and speeding reimbursement across a network that spans more than two dozen surgery centers built over 25 years of partnerships with hospitals and surgeons.
That kind of network-level rollout — one vendor, one coding engine, applied across dozens of facilities at once — is a different adoption pattern than the single-site pilots that defined ASC AI experiments a few years ago.
Early results ASCs are reporting
Becker’s ASC Review has reported that some centers are already pairing automated coding validation with broader revenue cycle automation. Allegiance Mobile Health’s CFO told the outlet the organization added nine new contracts in a year without adding revenue cycle headcount, and that automating eligibility checks and payment posting alongside coding validation helped lift the center’s first-month collection rate from the single digits into the 20% range. Those are self-reported, facility-specific numbers rather than an independent benchmark, but they illustrate why ASC finance leaders are paying attention.
What this means for coding accuracy and audit readiness
Faster claims are only a win if the codes are correct. Surgical coding carries real audit exposure — implant and device coding, modifier use, and bundling rules are frequent RADV and commercial payer audit targets — so an agentic system that moves a claim out the door faster but without a documented accuracy check just moves risk downstream. For ASC coding and compliance teams evaluating this category of tool, a few questions matter more than the speed pitch:
- Where does a human coder actually touch the claim, and is that touchpoint documented for audit purposes?
- Does the vendor publish an accuracy rate specific to surgical/ASC coding, not just evaluation-and-management coding?
- How does the system handle implant, device, and modifier edge cases that general-purpose coding AI tends to get wrong?
- Is there a defensible audit trail showing why a given code was assigned, not just that it was assigned?
- What happens on EHR integration failure — does the claim queue for human review, or does it submit on incomplete data?
None of that is an argument against adoption — the direction is clearly toward more automation in ASC coding, not less. It’s an argument for treating “agentic” as a workflow design question rather than a marketing label, especially in a specialty where a single miscoded implant line can trigger a payer audit.
The bigger pattern
ASCs adopting agentic coding tools at the network level, rather than facility by facility, suggests this category is moving past the pilot phase. For coding and compliance professionals, that makes it worth tracking which vendors are being deployed at scale, what accuracy data they’re willing to share, and how audit trails are built into the workflow — not just whether claims move faster. Expect more multi-facility ASC networks to follow the same path CSA and CBFS took: pick one coding engine, integrate it once with the EHR in use across the network, and roll it out site by site rather than negotiating a separate pilot at every location.
Keeping pace with that shift is easier with a platform built for it. Medikode’s automated medical coding platform is designed to pair the speed of agentic coding with the documented accuracy and audit trail that surgical and specialty coding demand.