Long before Ali Sadat built software for dental offices, the back office was where he did his homework. His mother worked at a practice in Vacaville, California, and through middle school and high school she kept him, as he later put it, on an “extremely tight leash.” The arrangement gave him an accidental education. A dental visit might happen in the operatory, but its financial afterlife unfolded among forms, attachments, phone calls, benefit details and claims.
Most children would remember the boredom. Sadat remembered the system. He saw how much invisible work sat behind a routine appointment and how easily uncertainty passed from an insurer to a front desk, then from the front desk to a family. Years later, after training in engineering and dentistry, he described those afternoons as an eye-opener. The professional knowledge came later. The irritation was already there.
This is the useful way to understand Retrace, the San Francisco company he founded in 2016. Its origin is less a flash of startup inspiration than a long exposure to an ordinary problem. Claims are slow because several parties must agree on what happened, what is covered, which evidence is required and when money should move. Each party has its own system. The friction lives in the handoffs.
A founder formed by the paperwork
Sadat’s connection to dentistry ran through his mother, but his path was not simply inherited. He built a deliberately mixed education. At the University of California, San Diego, he studied bioengineering and technology management. At the University of California, San Francisco, he trained in dentistry and oral and craniofacial sciences, earning his DDS in 2014. He also held clinical and research appointments at UCSF and Zuckerberg San Francisco General Hospital.
The combination matters. An engineer learns to decompose a process. A dentist learns what information has clinical meaning. A researcher learns to distrust an elegant answer until it survives contact with evidence. An operator learns that a correct decision arriving two months late can still be expensive. Sadat collected all four perspectives before turning the old back-office frustration into a company.
“My adolescent experience was a real eye-opener into today’s challenges.”Ali Sadat
He could also see a larger technological pattern. As an eighth grader, Sadat spent three days at UCSF while his mother took dental-assistant tests. He already knew that a UCSF professor had helped found Genentech. The detail is charming because it reveals an early habit: he paid attention to the bridge between research and application. In his eventual UCSF application essay, he wrote that his life’s goal was “to bridge the gap between the benchtop and the bedside.”
At UCSF, mentor Joel White gave that ambition a dental-data vocabulary. Sadat has called White “the original big data and dentistry guy,” crediting him with helping establish the intellectual ground for contemporary work in dental AI. The phrase “big data” has since been sanded down by overuse. Here it meant something concrete: enough structured examples to compare evidence, recognize patterns and make an administrative decision more consistent.
Where a claim loses its momentum
A dental claim looks simple from a distance. A practice provides a service, sends information to an insurer and gets paid. Up close, the line becomes a maze. Eligibility must be checked. Images and other documents may need to accompany the claim. The payer reviews the evidence against its rules. Missing or mismatched information causes another round trip. A patient waiting to understand a cost and a practice waiting for cash are both downstream of the same delay.
Retrace’s wager is that connectivity and machine learning can compress those handoffs. The company has described itself as a healthcare clearinghouse for dental practices and payers. Its products have addressed real-time eligibility, claim submission, review and payment. The aim is not to make the form prettier. It is to let the underlying information travel far enough, and clearly enough, for a decision to arrive while it is still useful.
The technical work reflects Sadat’s research background. He is listed as an inventor on granted patents assigned to Retrace Labs, including systems for converting information in dental images into text and for synthesizing dental imagery with generative adversarial networks. By April 2022, the company said it had filed more than 25 patents, with eight granted, and established more than 503 direct insurance connections.
Those numbers describe two different moats. Patents attempt to protect how the software interprets evidence. Connections determine where the result can go. In administrative technology, intelligence without a network can become an impressive demonstration stranded outside the workflow. A network without intelligence merely transports the old ambiguity faster. Retrace has tried to develop both at once.
When a workflow feels inevitable, draw every handoff. The opportunity often lives in the waiting between boxes.
Friends, family and a very exacting test group
Sadat launched Retrace with more than $1 million from friends and family. In November 2019, the company closed an $18 million Series A, with Intel Capital leading and SoftBank Ventures Asia supplying a substantial portion of the round. Institutional funding changed the resources available. It did not change the first test environment.
His mother and the staff at her dental practice reviewed the product. Their notes went beyond whether a button worked. They challenged terminology. They commented on colors and fonts. It is a small anecdote with a large product lesson: the people who carry a workflow know where software vocabulary becomes alien, where a harmless design choice creates hesitation and where a shortcut invented by an engineer adds work for everyone else.
Sadat’s mother had supplied the original observation deck and now supplied quality control. The loop from childhood to company became literal. A practice similar to the one that exposed the problem helped judge the proposed solution.
Make uncertainty visible, then remove it
Sadat’s public comments return to transparency. “Hopefully we’re bringing transparency to the claims process,” he once said. Elsewhere, he reduced the desired outcome to five words: “No surprise bills in the mail.” Both statements are modest by startup standards. Neither promises to reinvent dentistry. They describe a system in which the people involved know what will happen before the envelope arrives.
That restraint fits the work. Claims administration is governed by details, not slogans. An image must retain the information that matters. A model must be checked against evidence. A connection must deliver the right data to the right party. Sadat’s inclusion in a 2025 review draft for a proposed American Dental Association standard on quality assurance for dental cone-beam imaging is consistent with that evidence-first posture. The technology matters only if its output can be trusted.
The founder archetype here is neither pure clinician nor pure technologist. Sadat is a translator among specialists. He can recognize the practice employee’s complaint, the payer’s requirement, the engineer’s model and the researcher’s demand for validation as parts of one system. The breadth is useful because administrative failures rarely respect professional boundaries.
“To bridge the gap between the benchtop and the bedside.”The goal Sadat wrote in his UCSF application essay
Retrace effectively adds a third point to that bridge: the back office. The bench can produce a technique. The chair can produce clinical evidence. But the value still has to pass through systems that decide coverage, documentation and payment. Sadat’s company lives in that less-photogenic final mile.
There is a broader founder lesson in his path. Proximity to a problem is helpful, but proximity alone can turn into folklore: this is how the industry works, everyone knows it is messy, nothing can be done. Sadat paired proximity with range. Engineering let him model the process. Dentistry let him understand the evidence. Research taught him to test. Operating forced each idea through actual office routines.
The result is a career built around an old observation that never stopped being relevant. The administrative wait inside a dental practice looked small from the outside. From the back office, it touched every party. Sadat kept following that wait until it became a map, then a company. Retrace is the attempt to shorten the distance between evidence and answer, one handoff at a time.