Revenue Cycle AI Register, an example edition

Ostrander Health Finance

Revenue Cycle AI RegisterQ3 2026 · October 2026

Revenue Cycle AI Register · US health systemsData to September 30, 2026

Prior authorization leads new AI announcements for the first time

Health systems announced 79 revenue cycle AI deployments this quarter, 31 of them for prior authorization. Fewer than one in ten claimed results rests on an outside review or a staggered rollout

1The register by function

We added 79 deployments this quarter, bringing the register to 490. Coding and documentation remains the largest function overall, but prior authorization added almost twice as many new deployments. Two thirds of all deployments are now described as live rather than in pilot.¹

Table 1Announced deployments by revenue cycle functionnumber of deployments

Function	To date	New in Q3	Live	With claimed result
Coding and documentation	142	18	96	61
Prior authorization	128	31	71	44
Denials and appeals	97	15	64	38
Patient access and estimates	74	9	51	22
Collections and billing	49	6	37	15
All functions	490	79	319	180

Source: health system press releases, published case studies, earnings call transcripts and SEC filings, Ostrander Health Finance register

2Where new announcements are going

Prior authorization announcements rose from 19 to 31 in a quarter. Many systems tied them to the electronic prior authorization requirements that federal rules place on payers from 2026 and 2027, and to staffing shortages in authorization teams.

Figure 1New deployments announced per quarter, three largest functionsnumber of deployments

Source: Ostrander Health Finance register, by date of first public announcement

3How claimed results were measured

Of the 180 deployments with a claimed result, 97 compare a period before launch with a period after it, with no group of comparable work left untouched. Only 16 use an audited figure or a staggered or randomized rollout that separates the tool's effect from seasonality and payer behavior.²

Table 2Measurement basis of claimed resultsdeployments with a claimed result

Measurement basis	Deployments	Share
Before and after, no comparison group	97	54%
Vendor reported, no method stated	41	23%
Before and after, with comparison group	26	14%
Audited or independently reviewed	9	5%
Staggered or randomized rollout	7	4%
All claimed results	180	100%

Source: Ostrander Health Finance review of each announcement and case study

Figure 2Share of deployments with a claimed result, by functionpercent of announced deployments

Coding and documentation: 43%
Denials and appeals: 39%
Prior authorization: 34%
Collections and billing: 31%
Patient access and estimates: 30%

Source: Ostrander Health Finance register

4Against our KPI Benchmark

5What to watch next quarter

Whether prior authorization announcements hold their pace into the new year, how many coding deployments move from pilot to live, and the first linked results for patient estimate tools, which our benchmark participants begin reporting in the fourth quarter.

How this register is built

A deployment is counted when a health system or its vendor names the system, the function and the stage in a public source
Each claimed result is recorded with the metric, the period and the measurement basis the source states, or marked as unstated
KPI Benchmark links use only participants who consent, and no vendor sponsors or reviews the register

Approved for subscribersLaura WhitfieldHead of Revenue Cycle Research · October 9, 2026

¹ Live means the source describes the tool as in production use across at least one facility; all other deployments are counted as pilots.

² Shares are rounded to the nearest whole percent and sum to 100.