One miscounted unit per visit is a denial you find weeks later.
Units, modifiers, and visit limits slip across your locations, and you find out weeks later on the remittance. Relay builds AI employees wired into the systems you already run, so the work catches it overnight before it costs you, all HIPAA-compliant.
See where you're leakingWhere your pediatric physical therapy clinic is leaking revenue.
What you earn and where your hours go sit scattered across systems that do not talk to each other, multiplied by every location you run. Prior authorizations expire mid-plan-of-care. The 8-minute rule turns one miscounted unit into a denial. A patient's cumulative spend crosses the KX threshold and the modifier never gets added. Each gap sits in a different system until the remittance surfaces it, weeks late. AI employees pull it into one picture you can act on before the claim goes out, fully HIPAA-compliant.
AI employees go to work inside your back office.
In the first 2 to 3 weeks we map your operation and build your first working AI employee inside your stack, so you see it before we build the rest. Then our engineers embed in your clinic and build AI employees inside the systems you already run. They track authorizations per patient per location, flag KX modifier thresholds before a claim goes out, and reconcile billing overnight, so the errors that drive your denial rate get caught on the way out. Anything that matters routes back to your staff for approval first. A pediatric therapy client had all 7 of its locations running on a Relay custom build within 90 days.
Prior auth never expires mid-episode.
Each night the AI employee sweeps every active authorization across every location, compares remaining approved visits against scheduled appointments, and surfaces what is about to lapse. Staff see a reviewed queue in the morning, not a denial three weeks later. Plans of care due for renewal get flagged the same way, so a missed physician signature does not become a retroactive write-off.
Billing closes its own loop.
Timed CPT codes like 97110, 97112, and 97530 require combined-minute calculation across a full session. The AI employee runs that math on every claim before submission, flags unit mismatches and any 97014 billed beyond one unit, and matches each remittance back to its claim and note. Unbilled hours stop slipping, and coding errors get caught on the way out, not on the way back.
How we build it.
We start from the problem you feel, then build the fix on the systems you already run. Discovery and your first working AI employee take 2 to 3 weeks. The full build runs 8 to 12 weeks.
Start with a free 30-minute call
A short call about where the work is piling up and what that is costing you while it stays manual. No commitment, and you leave knowing where you would start.
Discovery and your first AI employee (weeks 1 to 3)
A few working sessions with your team. We map your operation end to end, every workflow across your locations, and find where the money leaks and what closing it is worth. You do not walk away with just a document. By the end of discovery we have built your first working AI employee on top of the systems you already run, so you see it pay off in your real setup before the full build starts.
The full build (8 to 12 weeks, start to finish)
We build the rest of the AI employees you mapped and wire them across every location. Nothing goes out until your team approves it, so you stay in control the whole way. One pediatric therapy client had all seven locations live within 90 days.
Common questions.
Does Relay replace our WebPT or CentralReach?
No. Relay builds on top of the systems you already run, including WebPT, CentralReach, and SimplePractice, rather than replacing them. The AI employees wire into your existing stack and automate the back-office work that sits between those systems. Your team keeps the tools they know.
How does Relay handle prior authorization tracking across multiple locations?
AI employees sweep every active authorization each night, compare remaining visits against the schedule, and surface what is about to lapse before a session gets billed to an expired auth. Each location runs the same check, and staff approve anything that needs action before it goes out.
What is the 8-minute rule and how does Relay prevent billing errors?
Timed CPT codes like 97110 and 97530 are billed in units using combined total minutes across all timed services in a session, not per code. The AI employee calculates units correctly on every claim and flags any session where the math does not hold, including untimed codes like 97014 that are always one unit regardless of duration. Errors get caught before submission, not on the remittance.
What happens when a patient hits the KX modifier threshold?
For 2026, the combined PT and SLP threshold is $2,480. Above it, claims require a KX modifier with documented medical necessity. The AI employee tracks each patient's cumulative spend against the threshold and flags the chart when documentation is needed before the next claim goes out, so the modifier is never missed at volume.
How long does it take to get all our locations running on Relay?
Discovery takes 2 to 3 weeks and ends with your first working AI employee live in your stack. The full build runs 8 to 12 weeks end to end. A pediatric therapy client had all 7 of its locations running on a Relay custom build within 90 days.
How does the engagement work after the build is done?
After the 8 to 12 week build, the engagement continues monthly. That keeps the AI employees maintained and updated as your payers, stack, and staffing change. A person on your team approves every action before it goes out, so the human-in-the-loop step never disappears.
See where the gap is between your denial rate and 5%.
The call is a free 30-minute working session. We map 3 back-office workflows where the AI employees would close the loop automatically, and you leave with a picture of where the billing is slipping and what it costs you each month. The longer the gap stays open, the more it compounds. No commitment to go further.