Jeffrey Morelli went from selling ABA session-note software to building a care readiness platform, and argues most denials are administrative, not clinical.
Key Takeaways
- His diagnostic question for ABA operators is a ratio. Morelli tells leaders they cannot call turnover a staffing shortage until they can say how many minutes of administrative work support one hour of clinical care. “Death by 1,000 clicks is still death,” he told a Boston panel in August.
- Silna is betting the money moves to the front of the revenue cycle. Morelli estimates 80 percent of what providers spend capturing revenue goes to claim submission, denial management, and reconciliation, while the workflows that cause most denials sit in the remaining 20 percent.
- Scale is the mechanism, and the scale figure is the company’s own. Morelli says Silna processes more ABA prior authorizations than any single ABA provider in the country, which lets it detect changes in payer criteria and enforce them before the next submission. Acuity could not independently verify the claim.
- He draws a hard line at clinical content. Silna flags treatment plans it expects to be rejected but does not propose changes to a plan of care, and the company puts about 65 percent of the authorizations that come back rejected or short of what was requested in the administrative category rather than the clinical one.
At a summit in Boston on August 6, on a panel titled “The Hidden Cost of Bad Tech: How Poor Implementation Drives Turnover in ABA,” the moderator put a question to a row of technology executives: how often does an ABA organization blame turnover on a staffing shortage when the real problem is the way the work is designed?
Jeffrey Morelli, the Co-Founder and Chief Executive Officer of Silna Health, answered by proposing a number that most of the room would not have. “Can you tell me the amount of minutes that’s required through technology and administration work to support one hour of clinical care?” he asked. “And if you don’t know that number, then you can’t confidently say that this is actually a staffing shortage.” Then he made the point in the vocabulary of the audience, most of whom were behavior analysts or sold to them. Clinicians are trained to read behavior as a function of environment, he said, and are slow to apply that lens to their own organizations. “Death by 1,000 clicks is still death. You just may not assign a certificate against it.”
It is a characteristic move: reframe a staffing complaint as a measurement failure, then supply the measurement. Silna sells the software that produces it.
From ABA Session Notes to a Care Readiness Platform
Morelli sat down with Acuity Media Network last week for a wide-ranging interview about the company, the second Silna executive to do so in as many weeks. More than a decade ago, he left school in Utah and moved to San Francisco to sell software to behavior analysts. The company was Nightingale, and the product was unglamorous: a way for clinicians to take session notes on a tablet rather than on paper. It was his first exposure to ABA and to the paperwork surrounding it, and the company was eventually sold. He describes it now with some distance. “In hindsight, it was essentially AI notating scribes for AI,” he said. “Right inefficiency, right pain point, wrong technology paradigm shift.”
The next decade was spent in an unrelated industry. Morelli joined Truework, an income and employment verification startup later acquired by Checkr, as its first sales hire, and rose to lead data partnerships. Sagar Jajoo, now Silna’s Co-Founder and Chief Operating Officer, was Truework’s first product manager. The third co-founder, Pavel Asparouhov, grew up in Morelli’s Salt Lake City neighborhood, left UC Berkeley after his sophomore year, and spent three years at the fintech company Ramp building the team behind its payments and workflow automation products. The three started Silna in New York in August 2023, with Accel leading a $5 million seed round that month. In March 2025 the company announced a $22 million Series A co-led by Accel and Bain Capital Ventures, bringing total funding to $27 million, with angel investment from the founders of Ramp, Opendoor, Truework, Eight Sleep, and Speak. It has not announced a round since.
The idea came from a conversation with a relative who ran hospital systems and from watching his mother, a nurse for 40 years, spend the last stretch of her career frustrated. What struck Morelli was the direction of the industry’s technology response. “Everyone in healthcare from a technology perspective seems to be obsessed with how do you make a broken process more efficient,” he said. “What is the best way to fight a denied claim? And it felt like no one was asking the question of how do you make it so the system never breaks down in the first place.”
Why Silna Calls Itself a Care Readiness Platform Instead of Front-End RCM
The distinction rests on an arithmetic claim. By Morelli’s estimate, roughly 80 percent of what a provider spends capturing revenue goes to the middle and back of the cycle: submitting claims, managing denials, reconciling payments. Only about 20 percent sits at the front, in eligibility checks, benefit verification, and prior authorization, which is also where he says most denials originate. Acuity has reported separately on how much of that front-end work in ABA still runs through phone calls rather than electronic transactions.
For a while the company described itself as front-end revenue cycle management. Morelli came to regard that as a failure of nerve. “What I’m telling you is what we’re not,” he told Acuity. “We’re not RCM. We’re the front side of it.” So Silna convened a customer advisory board and asked users what the term evoked. One word came back repeatedly: reactive.
The alternative he settled on is care readiness. “Readiness should evoke this emotion that says you are confident that this patient is ready to receive care, that the payer thinks that care is medically necessary,” he said. He is explicit that he is trying to establish a category rather than a tagline, and reaches outside healthcare for the precedent: customer success management, he noted, was not a recognized function until Gainsight argued that it should be one. He has not trademarked the phrase and says he does not intend to. “I think there will be other care readiness platform providers out there, and when they are, I’ll know that we’ve been successful, because folks are trying to copy Silna.”
Naming a category is slower than deciding on one. On the Boston panel a month earlier, given 30 seconds to introduce himself to a room of operators, Morelli described Silna as “all front-end RCM of what’s required to financially clear a patient for care.” The older shorthand is still the one that lands fastest.
How Silna’s ABA Prior Authorization Volume Feeds Its Document Validation Engine
Morelli’s argument for why general-purpose vendors fail in ABA is specific rather than rhetorical, and he makes it the same way in both settings. Benefit verification in many healthcare specialties can be handled electronically, through a payer integration returning an EDI 271 response. In ABA the benefit is usually carved out. Unless the payer is Medicaid or an MCO, he told the panel, obtaining ABA benefits requires a phone call almost all of the time. Vendors arrive claiming they are good at verification of benefits, he said, without understanding how often behavioral health is carved out to a third-party administrator. The result, in his account, is a market full of tools rather than solutions: “They make you more efficient at certain parts of the job, but you still need folks to handle this themselves.”
The product that follows is called document validation, and it attempts to determine the outcome of an authorization before the authorization is submitted. A provider uploads a treatment plan, Silna runs it against the payer criteria it has encoded, and the system flags where it expects a problem. The provider can submit anyway.
What makes that possible, in Morelli’s account, is volume. “Right now, Silna has more patients being served in Silna than any individual provider,” he said. “We are larger than the largest ABA provider in the country when it comes to prior authorizations.” Asked for something behind the comparison, Silna said the top seven ABA providers on its platform account for roughly 12,000 active patients between them. Acuity could not independently verify the comparison itself, and the company has not published a total ABA authorization count. Taken at face value it would place Silna’s volume above that of the largest ABA platforms in the market, none of which are software companies. The company does publish broader figures: it says it works with more than 1,000 payers across 50 states and supports more than 250,000 patients, against the 50,000 it cited at launch in March 2025. That total spans physical, occupational, and speech therapy, behavioral health, and post-acute care as well as ABA, so it does not isolate ABA.
The claim matters because of what he says it produces. Silna reviews authorization outcomes across its client base over rolling 30, 60, and 90 day windows, looking for shifts in what payers accept, then pushes any new rule to the front end for every client facing that payer. “The more authorizations we handle, the better the product gets,” Morelli said. “The better the product gets, the more customers that we can sign on.”
Some of what surfaces is substantive, such as North Carolina moving reauthorization for treatment plans above 16 hours a week from six months to three, part of a broader tightening of that state’s ABA benefit. Some of it is not. Morelli described one state Medicaid program, which he declined to name, that produced a sharp rise in denials last year over the formatting of graphs. “It had nothing to do with the information in the graph or the other 40 pages of documentation,” he said. “It was purely if your format on the graph was wrong, the thing would be denied and kicked back.” Silna’s response was to push a formatting rule to every client submitting to that program.
Jajoo, who spoke with Acuity separately in early September, put numbers to the same problem. Silna says 35 percent of the ABA prior authorization requests it submitted in North Carolina between June 1 and August 31 were not fully approved on the first pass, against 27 percent across its nationwide book, on roughly 1,500 North Carolina authorizations out of about 9,000 processed nationally. The category covers straight denials, partial approvals, and requests for additional information. The figures reflect the company’s own client base rather than the market, and Acuity could not independently verify them.
Asked by Acuity where a model makes a decision and whether Silna touches clinical content, Morelli drew a firm line. “Only completeness,” he said. “We do not change anything on the clinical documentation side.” The system flags patterns it expects will draw a rejection but does not propose changes to a plan of care, “because ultimately we believe the provider should be the expert in determining the plan of care for that patient.” What it does address is administrative: rendering provider details, dates of birth, submission mechanics. Silna says that split holds across what it sees. Of the authorizations that come back rejected or approved for less than what was requested, the company puts roughly 65 percent in the administrative category: member eligibility, an invalid order, an out-of-network provider, a service that is not a covered benefit, or an overlapping authorization already on file. The remaining 35 percent are clinical, including a requested peer-to-peer review, a medical necessity question, or insufficient documentation. Not every one ends in a denial. The distinction is worth watching as federal audits of ABA documentation continue to work through state programs.
Silna publishes customer case studies, and Morelli cited two. One, with the multi-state provider ABS Kids, involves revenue the organization had been writing off because it could not detect when patients changed plans; Silna’s published version says ABS Kids completes authorizations six times faster and attributes more than $1 million in annual financial impact to the platform. Another, with a behavioral health and post-acute provider, reports clinicians seeing 8 percent more patients because intake moved faster. Both are company materials rather than independent assessments.
What Morelli Tells ABA Providers About Buying and Implementing Technology
The Boston panel was where he talked about how software fails, and his answers there were less about products than about the organizations installing them. Implementation, he argued, is the wrong noun. “Implementation plus calibration,” he said. “Implementation feels far too static.” He is similarly hard on go-live dates: “Implementation should never have a countdown. It should have a goal.” Hiring a head of marketing does not solve marketing on the person’s start date, and a system does not solve a workflow on the day it turns on. The questions operators should ask before signing follow from that framing.
To distinguish real failure from ordinary friction he offered a rule: questions are noise, workarounds are signal. Staff asking how to find something are uncomfortable; staff who have built three new spreadsheets to manage the tool you just bought are telling you the system failed. His illustration was Facebook’s News Feed, which launched in September 2006 to immediate revolt. Facebook had roughly 9.5 million users, and the largest protest group, organized on Facebook itself, passed 700,000 members within days. The company held. “You always need to decipher the noise from the behavior,” Morelli said, because the behavior showed users spending more time on the product than ever.
He also cited a 1990 Stanford study on the difficulty of explaining anything you already understand. Elizabeth Newton had participants tap the rhythm of well-known songs on a table while others tried to name the tune; the tappers predicted listeners would get about half right, and of 120 songs listeners identified three. Morelli said he raises it with every new Silna employee. “Once you have context in your brain, you can’t imagine not having that context in your brain,” he said, and the conclusion he draws is that leaders have to make it safe to ask the question that sounds stupid.
The most useful thing he said on the panel was about his own company. Silna decided last year that its onboarding process was too slow, set a north star of speed to go live and a counter-metric of customer experience, and built what it believed was a rigorous new process. “Only to realize two weeks in that we had overbuilt it,” he said. Too many handoffs, too many people, and the software they had chosen was the wrong one. They switched vendors, cut two internal handoffs, and halved implementation time. He described the salvage as a function of having agreed on the goal in advance: every meeting opened by restating the north star before anyone reviewed the process.
His advice on evaluating vendors is unsentimental about the category he sells into. “The landscape is riddled with folks who overpromise and underdeliver, have really shiny demos, have really beautiful websites,” he said, “and the reality is you can just do it better internally with your staff.” The pressure test he recommends is to ask a vendor who the product is not good for and when it has failed. On automation specifically, he proposed a sorting rule: put it where a mistake is tolerable, keep it away from family communication and clinical outcomes, and in between insist on systems that let a human intervene mid-process. “It’s a lot scarier when you have a black box.”
He applies the same logic to his own build-versus-buy decisions, and volunteered a case where he got it wrong in the profitable direction. Silna bought a vendor to handle payer communication because engineering time was scarce. Nine months later, at scale, the tool proved too limited, and his team replaced it internally in about two weeks. It is the kind of calculation that separates software companies from the RCM services firms that compete with them for the same ABA budget.
What ABA’s Outcome Measurement Problem Means for Payer Scrutiny
Morelli does not describe the current environment as a crisis, though he does not dismiss the word. “Depending on the state you’re operating in, the stage of your organization, I think for some folks this is certainly a crisis that is developing,” he told Acuity. On a longer horizon he reads it as a growing pain in an industry reaching critical mass.
His explanation for the timing is structural. Autism therapy, he said, has carried one of the highest rates of third-party administration in healthcare, because payers did not know how to underwrite it and outsourced the judgment. That changes once the population is large enough to matter to a network. Autism prevalence among eight-year-olds now stands at about one in 31, according to CDC surveillance data published in April 2025 covering 2022, up from one in 36 two years earlier. “We’re paying out hundreds of millions of dollars a year for these services,” Morelli said, voicing the payer. “How do we think about outcomes? How do we prove that care is medically necessary?” It is the same arithmetic that has drawn private capital into behavioral health and state legislatures into the benefit.
He offered two precedents. Virtual behavioral health expanded during COVID, drew payer attention as costs rose, and in his account survived by proving outcomes well enough that payers came to see the spending as reducing long-term cost. Dialysis was fragmented, drew scrutiny, consolidated, and now generates comparatively little friction. Whether either maps onto ABA is unsettled, and he did not claim certainty.
What he did claim is that ABA has given payers nothing to underwrite against. At a session Silna presented at the Autism Investor Summit West in May, according to the event agenda, he made the point with a thought experiment: hand 20 people in the room a piece of paper and ask how outcomes in ABA should be measured, and you get 20 answers. “That is problematic, because then as a payer you say, wait, I don’t know how to think about outcomes, because you all don’t know how to think about outcomes.” He called for a coalition to settle the question without specifying who would convene it. The measurement debate remains unresolved among clinicians, which is part of his point.
On fraud he was careful to scope it, describing bad actors as a small sliver of the industry that creates headwinds for everyone else, and then explaining why the sliver exists. A fraudulent actor would look for two things, he said: a high cost of care, because low-value sessions are not worth the effort, and ambiguity in how outcomes are judged. “Unfortunately, at the intersection of that is ABA.”
That framing produces the most speculative thing he said, and he flagged it as speculation before finishing the sentence. Rather than treating authorization as friction to be reduced, he wondered what happens if it disappears from view. “What if the process was so seamless that an authorization could be submitted to an insurance company every 30 days, but the provider never knew it?” Raw session notes would assemble into a treatment plan and the submission would be automatic. Then he cut himself off. “I’m not arguing we should get an authorization every 30 days. We should not.” The point, he said, is that if review were invisible, its frequency would stop being what providers fight about, and authorization would work as a barrier to bad actors rather than a tax on everyone else.
It is a vision with obvious commercial logic attached, since the company describing a future in which submissions stop hurting is the one automating the submissions. Morelli did not pretend otherwise, and when asked what he actually optimizes for, he gave an answer with no product in it. His north star, he said, is a reduction in pajama time: whether clinicians have stopped finishing their documentation at home. “Everything in healthcare, in my mind, comes down to trust.”
