Tally, an AI-native RCM platform built for multi-site specialty practices, has announced a strategic partnership with Liminal — a Founders' Studio backed by Temasek. The combination pairs agentic AI billing with deep build capability to go after one of the most under-served segments in healthcare revenue cycle: the specialty group that's too large to ignore and too small for hospital-focused platforms to care about.
The announcement landed July 23, 2026. Tally was founded by Peter van Riper — a 25-year healthcare executive who previously scaled and exited a multi-site dental company at a $90M valuation — alongside co-founder Satish Cheema, who helped architect the platform behind a kidney care company's growth from $25M to $3.3B in enterprise value. These aren't first-time founders. They've run the operations they're now automating.
Share of providers reporting denial rates above 10% — up from 30% in 2022. The denial problem is getting worse, not better.
Most RCM AI vendors sell a worklist. They surface the denials, flag the underpayments, identify the prior auth risks — and then hand it back to your billing team to fight. That model is running into its limits. Denial rates are still climbing. Staffing gaps are widening. And the offshore labor arbitrage that propped up the old model is under structural pressure from the same AI that's supposed to fix it.
Tally's positioning is different: they run the full cycle. Insurance verification, prior authorization, claim submission, denial management, underpayment review, payment posting — agentic AI executes the workflow, with embedded specialists handling the exceptions. The platform doesn't advise. It works.
The logic behind this matters. If you only advise, your data is limited to what your clients share. If you operate the billing, you own the outcomes data — and every claim becomes a training signal.
"Every claim we work teaches the system how each payer really behaves," said Peter van Riper. Tally's proprietary dataset maps how individual payers actually adjudicate claims across procedure codes, documentation types, and reimbursement rules. That's not something a pure-software tool can replicate without running the billing itself.
The hospital-focused RCM AI platforms have dominated investment and attention for the past three years. Multi-site specialty operators — dental groups, PT practices, behavioral health, surgical centers — have been a rounding error in that funding story despite representing enormous claim volume.
The reasons aren't hard to find. Specialty billing is fragmented by vertical. Payer behavior varies significantly across specialties. The practices themselves are smaller and harder to land as enterprise customers. It's messier than health systems, and the contract sizes are smaller on a per-account basis.
But the math is changing. Median hospital operating margins now sit below 3%. The share of providers reporting denial rates above 10% jumped from 30% in 2022 to 41% in 2025. And 63% of providers report RCM staffing gaps. Those pressures hit specialty groups first — they have less capital, smaller billing teams, and less leverage with payers than health systems. The gap between what they need and what's available for them is widening.
Liminal's involvement is meaningful on the build side. Their portfolio includes co-founders of Twitter, Block, Pinterest, and Medium, and repeat founders with multiple exits. The firm is funded by Temasek, the Singapore-based global investment company. This is a well-resourced bet, not a scrappy seed-stage play.
The announcement is light on specifics that would let you evaluate execution risk. A few things worth tracking as Tally scales:
If you're running RCM for a multi-site dental, PT, or behavioral health group, Tally is worth watching. The end-to-end model is exactly what most specialty groups need and can't build internally. The risk is execution — running billing operations at scale is harder than selling software. Watch for early case study data before committing.
Tally is making a credible bet on a real gap. Specialty RCM is large, fragmented, and structurally under-served by the current generation of hospital-focused AI platforms. The Payer Behavior Graph is a genuine data moat — if the team executes and claim volume scales.
The risk is operational complexity. Running actual billing is harder than selling software to billing teams. The embedded specialist model adds overhead. Building payer intelligence across dozens of specialty verticals takes time.
But the market is there, the founders have lived the problem, and the Liminal backing is serious capital behind serious builders. If they can prove clean-claim rate improvement and denial reduction by specialty, this becomes a margin and labor story for every multi-site specialty operator in the country.
Daily coverage of payor policy, denials, and vendor moves — written for people who live in revenue cycle.