Campus to Commure: Boston University

Patient scheduling software gives patients a way to book, confirm, and change appointments on their own, and gives the health system a rules engine that keeps every booking accurate whether it started online, by phone, or through an AI agent. In an August 2026 MGMA Stat poll, 32% of medical groups reported higher no-show rates year to date than in 2025, while 58% saw no change and only 10% saw an improvement.

That poll arrived after a year in which practice leaders had already made scheduling a priority. When MGMA asked in December 2025 what leaders would focus on in 2026, the answers split almost evenly across no-shows (27%), online scheduling (24%), phone access (22%), and wait times (21%). All four are scheduling problems, and a health system that fixes them at the booking step fixes them for the intake and follow-up steps too.

What patient scheduling software does for a health system

Patient scheduling software is the system that turns provider templates, visit types, and business rules into bookable appointments, then lets patients and staff fill them from any channel. At a health system, that means one ruleset governing thousands of providers across dozens of locations, with every booking written back to the EHR in real time so the schedule the patient sees matches the schedule the clinic runs on.

The distinction that matters most is between a request form and true scheduling. A request form collects a preferred time and hands it to a staff member who calls back, which saves the patient a phone call by creating one for your staff.

True scheduling shows live availability, applies the provider's rules, books the slot, and confirms it while the patient is still on the page.

Scheduling sits at the front of the digital front door, and it's the first place a patient learns whether your organization is easy to deal with. If booking takes three transfers and a callback, the intake, communication, and care navigation that follow already start off on the wrong foot.

Core features of patient scheduling software

Patient self-scheduling and online booking

Patient self-scheduling software lets a patient pick a provider, location, visit type, and time from live availability without staff involvement. The best implementations don't require a portal login, because every account-creation step drops completion rates, and they support new patients as well as established ones.

Online patient scheduling should also hand off cleanly to intake, so the same platform sends patient intake software forms, insurance capture, and prep instructions the moment a booking lands.

Rules-based provider and visit matching

A scheduling rules engine encodes what your schedulers carry in their heads, including which providers see new patients, how long each visit type runs, which payers a provider accepts, and which conditions need a specific specialist. It applies those rules identically online, in the access center, and through voice AI, so a patient gets the same correct answer regardless of how they asked.

At health system scale, someone has to own the templates, someone has to approve changes, and the software has to make those changes take effect everywhere at once.

Reminders, confirmation, and rescheduling

Automated reminders by SMS, email, and voice are standard, and the reply path is what separates modern scheduling from a reminder tool. A patient who can confirm, cancel, or reschedule by texting back keeps the slot in play, which is the same two-way loop that has turned patient communication software into a conversational layer.

Waitlist and slot backfill

When a patient cancels, the software should offer that slot to the next eligible patient automatically, filtered by visit type, provider, and location. Manual waitlists fail because the front desk has the least time to work them on the busiest days, and automated backfill is one of the most direct ways to reduce no-shows and the empty slots they leave behind.

The 2025 MGMA DataDive reported a 19.95% appointment cancellation rate in its 2024 single-specialty aggregate, with only 27.40% of those cancelled visits rescheduled within 30 days. MGMA cautions against treating the cancellation figure as a universal target, but the reschedule figure carries no such caveat. Roughly 3 in 4 cancelled visits never came back within a month.

Integration is where patient appointment scheduling software succeeds or fails

Integration depth decides whether scheduling software removes work or relocates it, and vendors describe very different things when they say "integrated." The shallowest version reads availability from the EHR and writes back a confirmed time, which breaks the moment a health system needs visit-type logic the EHR doesn't expose. The middle level is bidirectional and real-time in both directions, and it's the minimum for a health system because a lag of even a few minutes produces double bookings on high-demand days.

The deepest level treats booking as the trigger for everything downstream. Scheduling a visit starts eligibility verification, sends intake, queues prep instructions, and sets up follow-up without anyone initiating those steps, and checking coverage at that moment catches the appointments that would otherwise become denied claims.

For a health system, the other question is where the scheduling rules live. If they live in the EHR, you'll maintain them in several places and watch them diverge at acquired sites. If they live in the scheduling platform and execute against each EHR, you maintain them once.

Healthcare scheduling software systems across every channel

Patients book on whatever channel is in front of them, so the scheduling system has to be equally correct online, in the access center, and through voice AI. Phones still carry a large share of that volume. In a March 2026 MGMA Stat poll, practice leaders named scheduling as 31% of their most time-intensive phone work, second only to eligibility and prior authorization at 45%.

The enterprise version gives access center agents the same guided, rules-based booking flow the patient sees online, and hands routine calls to AI call center agents that book, confirm, reschedule, or cancel against live EHR data at any hour. One community health center working with Commure was able to expand call center capacity by roughly 800 labor hours per 10,000 calls, with 30% to 80% of inbound calls resolved without staff intervention.

What to measure once patient scheduling software is live

Scheduling metrics only mean something when you know which direction is good, so each one here states its direction first.

Third next available appointment. Lower is better. This is the standard access measure, counting days until the third open slot for a given provider or visit type, because the first and second openings are often cancellation artifacts. Track it by specialty, since averages hide the departments where access is failing.

Self-scheduling adoption. Higher is better. Measure the share of total bookings made without staff, and break it out by new versus established patients. Low adoption almost always traces to a login requirement, a request-form workflow, or provider templates that were never opened to online booking.

No-show rate and cancellation rate, tracked separately. Lower is better for both, but they mean different things. A cancellation gives you notice and a chance to refill, while a no-show usually doesn't. Pair them with the share of cancelled visits rescheduled within 30 days, which shows whether the patient is coming back or leaking out of the system.

Schedule fill rate (higher is better) and access center call abandonment (lower is better) round out the picture, and both should move within the first quarter if backfill and AI agents are doing their jobs. Read all of these alongside your patient experience scores, and map them to the access stage of the patient journey so improvements here can be traced to the intake and follow-up stages that depend on them.

How to evaluate the best patient scheduling software for a health system

The best patient scheduling software for a health system is the one that runs your rules correctly in every channel, writes to every EHR you own in real time, and triggers the rest of the pre-visit workflow without staff. Use these questions in demos and RFPs.

• Does the patient see live availability and get an immediate confirmation, or is this a request form with a callback?
• Can a new patient book without creating an account?
• Where do scheduling rules live, and do changes propagate to online, access center, and voice channels at the same time?
• Which EHRs are supported with bidirectional, real-time integration, and how are acquired sites on a different EHR handled?
• Does booking trigger eligibility verification, intake, and prep automatically, and does the waitlist refill open slots without staff?
• Is voice AI included, and does it book against the same rules and EHR data as the web experience?

Any vendor should be able to answer all 6 in a single demo.

How Commure Engage approaches patient scheduling

Commure Engage treats scheduling as one connected workflow with intake, communication, and care navigation. Patients book and manage appointments by text or voice without an app or portal, AI call center agents handle scheduling calls against live EHR data around the clock, and automated care pathways carry the patient from booking through prep and follow-up.

The results show up in the metrics above. Health systems using Engage have seen a 6.5x increase in outreach capacity for scheduling and coordination, a 31% reduction in inbound call volume, and 80%+ of patient inquiries resolved automatically. At Yale New Haven Health, bilingual pre-appointment texting for breast imaging cut no-show and same-day cancellation rates from 13% to 6%, a 54% reduction.

If your access metrics are stuck, the fastest way to find out what's fixable is to walk through them with our team.
‍

Explore Commure Engage

Frequently asked questions

What is the difference between patient scheduling software and a patient portal?

Patient scheduling software books appointments against live provider availability and business rules from any channel, including text, voice, and web, usually without a login. A patient portal is a logged-in account tied to one EHR that may include a scheduling request feature. Portals cover established patients who enroll, while scheduling software also covers new patients and the majority who never log in.

How does patient self-scheduling software work?

Patient self-scheduling software shows patients real-time availability filtered by the health system's rules for provider, visit type, location, and payer, then books and confirms the appointment instantly and writes it to the EHR. The patient can later confirm, cancel, or reschedule by replying to a reminder. Systems that require account creation or route requests to staff for approval see far lower adoption.

Does online patient scheduling reduce no-shows?

Online scheduling reduces no-shows when it's paired with two-way reminders and automated backfill, because patients who chose their own time and can reschedule by text are less likely to silently skip. Scheduling alone won't fix no-shows driven by cost or transportation. At Yale New Haven Health, pre-appointment texting through Commure Engage cut no-show and same-day cancellation rates from 13% to 6%.

What is third next available appointment and why does it matter?

Third next available appointment is the number of days until the third open slot for a provider or visit type, and lower is better. It's the standard access measure because the first and second openings are often cancellations that misrepresent true availability. Health systems track it by specialty to find where access is failing and to measure whether scheduling changes worked.

Can patient scheduling software integrate with multiple EHRs?

Yes, and for a health system this is the deciding requirement. Enterprise patient scheduling software maintains one set of scheduling rules in the platform and executes them against each EHR through bidirectional, real-time integration, so acquired sites on a different EHR follow the same booking logic. Ask vendors which EHRs are supported and whether writeback is immediate.

Is patient scheduling software HIPAA compliant?

Patient scheduling software handles protected health information, so the vendor must sign a business associate agreement and the platform must encrypt data, control access by role, and log activity. Confirm that the agreement covers every module you'll use, including intake forms and payment, since some vendors exclude parts of the platform or reserve the agreement for higher pricing tiers.

What should a health system look for in patient scheduling software?

A health system should look for live self-scheduling without a login, a single rules engine that runs identically online, in the access center, and through voice AI, bidirectional real-time integration with every EHR in its footprint, automated waitlist backfill, and reporting on third next available and fill rate by specialty. Platforms that connect scheduling to intake and follow-up remove the most staff work.

Only 9% of healthcare finance leaders are very confident their organization captures all the revenue it's owed, and 60% say they need more insight into why claims get denied, according to a 2025 HFMA survey of 272 executives. The data exists somewhere in the health system, but nobody can see it in one place in a timely manner.

Revenue cycle analytics is how a health system closes that gap. It turns the claims, remittance, coding, and registration data already flowing through your systems into a picture of where money is stalling, where it's leaking, and what to fix first.

What is revenue cycle analytics?

Revenue cycle analytics is the practice of collecting financial and operational data from every stage of the revenue cycle, from scheduling and eligibility through coding, claim submission, adjudication, and payment, then analyzing it to find the root causes of lost or delayed revenue. In a hospital setting, it answers questions like which payers deny most often and why, where charges go missing between the exam room and the bill, and how long cash sits in accounts receivable by service line.

Healthcare revenue cycle analytics differs from general business intelligence in the data it works with. Every encounter generates a claim, a remittance with CARC and RARC codes explaining what the payer did, a documentation trail, and a set of timestamps. Revenue cycle analytics reads across those artifacts, so a denial spike at one facility can be traced back to a registration workflow or a documentation gap rather than logged as a payer problem.

The output is a shared, current view of revenue cycle management performance that finance, operations, and clinical leaders can act on together.

Why health systems can't get one number

The average health system runs on 150 or more software solutions, a figure Commure's SVP of Commercial Operations cited when connecting the exam room to the balance sheet. Revenue cycle data is scattered across the EHR, one or more billing systems, a clearinghouse, payer portals, and whatever spreadsheets each department maintains. Ask 3 teams for the denial rate, and you'll get 3 answers, because each team pulled from a different source with a different definition.

Hospital revenue cycle analytics has to solve that problem before it can solve anything else. The work is normalization, meaning every system's version of a claim, a denial, or a write-off gets mapped to one definition so numbers are the same across facilities and in the audited financials. Multi-hospital systems that grew through acquisition feel this most, since a legacy EHR at one campus and a different billing platform at another can make a system-level denial rate impossible to calculate without weeks of manual reconciliation.

Native EHR reporting rarely gets there on its own, which is why so many systems layer a dedicated analytics platform on top. The EHR sees what happened inside the EHR. It doesn't see what the payer did after the claim left, or what the billing system did when the remittance came back.

Revenue cycle analytics metrics that matter

HFMA's MAP Keys define 29 standardized revenue cycle KPIs with set formulas and data sources. Each metric below notes which direction is good, since a rising number is a win for some and a warning for others.

Denial rate (lower is better). Denied claims divided by total claims submitted, usually measured at first submission. When Kaufman Hall surveyed 103 hospital and health system leaders in 2025, 44% named high denial rates and administrative burden as their top managed-care challenge, ahead of reimbursement rates.

Analytics earns its keep here by breaking denials down by payer, reason code, facility, and service line so the fix lands on the actual cause. Our post on denial management covers the full process.

Clean claim rate (higher is better). Claims that pass payer edits without manual intervention divided by total claims. A falling clean claim rate points upstream to registration, eligibility, or coding.

First-pass yield (higher is better). Claims paid in full on first submission divided by total claims. This is the single best indicator that the front and middle of the revenue cycle are working.

Days in A/R (lower is better). Total accounts receivable divided by average daily net patient revenue. Watch it by payer and aging bucket, since a healthy blended number can hide one payer sitting at 90 days.

Net collection rate (higher is better). Payments received divided by payments contractually owed after adjustments. This is the metric that catches underpayments, which never show up as denials.

Charge lag (lower is better). Days between date of service and charge entry. Long lag drives timely-filing denials and is the earliest signal that charge capture is breaking down.

Cost to collect (lower is better). Total revenue cycle cost divided by cash collected. Every manual rework loop, appeal, and phone call to a payer shows up here.

Most analytics programs stop at the back end. The mid-cycle deserves the same treatment, because clinical documentation improvement gaps and coding variance are where revenue integrity problems start, weeks before they surface as a denial or a write-off.

From dashboards to recovered revenue

A dashboard that shows last month's denial rate is descriptive analytics. It tells you what happened. Predictive analytics uses claim history and payer behavior to flag which claims in the current queue are likely to be denied or underpaid before they go out.

The third tier, and the one that can meaningfully change financial results, is analytics wired directly to action, where the flagged claim gets corrected, the missing charge gets captured, or the appeal gets drafted without a person exporting a report and building a work list.

That last step is where most revenue cycle analytics software falls short. It surfaces the insight and hands the work back to a team that's already stretched, which is why 60% of finance leaders in the HFMA survey still say they lack root-cause visibility even though nearly all of them own reporting tools.

A 400-bed New York health system ran into exactly this. Its MEDITECH EHR and Athena billing system weren't integrated in a way that let leadership track documentation through to charge entry, so charges were going missing with no report to catch them.

After deploying Commure's Charge Note Reconciliation, which compares clinical notes to billing records and automatically flags documented-but-unbilled services, average monthly charges rose 20% (from $7.5M to $9.4M), collections rose 21%, and timely-filing denials fell 53%. The Director of Billing and Revenue Cycle Management described the change as finally having visibility into what physicians were doing.

How to evaluate revenue cycle analytics software

Start with coverage. Does the platform pull from every system in the revenue cycle, including the billing platform and clearinghouse, or only from the EHR? Does it normalize that data to one set of definitions so a system-level KPI ties out across facilities?

Then test depth and follow-through. Can you drill from an executive KPI down to the individual claim, payer, and reason code in the same tool? And when it finds a problem, does it act on it or hand you a list?

The last question separates a reporting layer from a revenue engine. Our guide to AI medical billing software goes deeper on what "act versus flag" looks like in practice.

Analytics that recover revenue with Commure RCM

Commure RCM connects clinical documentation, coding, charge capture, and claims into one data layer, so the analytics don't stop at the dashboard. Charge Note Reconciliation surfaces unbilled services automatically, denial tooling flags at-risk claims before submission, and payer-level reporting ties out to your financials because it's built on the same records.

If your team can name the denial rate but can't say with confidence which workflow is causing it, that's the gap analytics should close.
‍

Discover Commure RCM

‍

Frequently asked questions

What is revenue cycle analytics in healthcare?

Revenue cycle analytics in healthcare is the use of claims, remittance, coding, and registration data to measure and improve how a provider organization gets paid. It identifies where revenue is denied, delayed, underpaid, or never billed, and traces each problem to its root cause so teams can fix the workflow rather than rework individual claims.

How is revenue cycle analytics different from revenue cycle management?

Revenue cycle management is the full set of processes that turn a patient encounter into payment, from scheduling and eligibility through coding, billing, and collections. Revenue cycle analytics is the measurement layer on top of those processes, showing where the work is failing, quantifying the cost, and pointing to what to fix first.

Why isn't EHR reporting enough for revenue cycle analytics?

EHR reporting only sees data inside the EHR. It misses what the payer did after the claim left and what the billing system did with the remittance. In a 2024 KLAS report, health systems said they adopted third-party revenue cycle analytics because EHR tools couldn't aggregate data from multiple sources or support the customized reporting leaders needed.

How does predictive analytics reduce claim denials?

Predictive analytics scores each claim before submission using historical denial patterns by payer, procedure, diagnosis, and provider, then flags the ones likely to be denied so staff can correct them first. Because most denials trace to a small number of repeat causes, catching them pre-submission raises first-pass yield and cuts the rework that drives cost to collect.

What should a health system look for in revenue cycle analytics software?

Look for a platform that ingests data from every revenue cycle system, normalizes it to one set of KPI definitions, lets you drill from a system-level metric to an individual claim, and acts on what it finds rather than exporting a work list. Platforms like Commure RCM pair the analytics with automation so the flagged problem gets fixed on the same encounter.

Justin Paquette flew to Dubai with three bags: one carry-on with his passport and work laptop, one checked bag with clothes for two weeks, and a second checked bag packed full of Commure merch. A hospital group in the UAE had just started a 60-day pilot of Commure's Ambient AI, and he landed as optimistic as he had ever been about a customer. After more than a year in the making, he told himself the whole flight over that the next two weeks were going to move the needle, that this was the last one percent the customer would remember when the pilot ended and it came time to sign.

He spent his days on site with providers and his nights in the hotel with his laptop open. Providers at the site moved between English and Arabic inside a single visit, so Justin worked with engineering on how the system detects the language of each recording, building the change to hold across every language Ambient supports rather than solve for one customer. He drafted and tested a second change the same week, tuning provider activity logic, and engineering reviewed it and pushed it live for the customer inside a week. Two days after the language change shipped, notes across the site were generating 20% faster.

That contrast, between the suitcase full of merch and the hotel room with a laptop open, is more or less the job. Justin is a Sales Engineer at Commure, a little under two years into the company across two roles and almost a year on the sales engineering team.

His path here did not start in software. He grew up in central Massachusetts and went to a vocational  high school, where he studied plumbing. During his senior year, he was placed in a co-op program where every other week he reported to a local plumbing shop at 6 am, and spent the day building commercial apartments. He made $10.25 an hour and thought it was a fortune, and he wasn’t sure he would go to college at all.

He ended up in the computer engineering program at UNH, then spent the first years of his career working on point-to-point radio and IoT networks, first in San Antonio and then back in the Northeast. By his own description, he does not think like the stereotypical Silicon Valley startup person, and he wants candidates to know that the background doesn't matter as much as what you do in the room. If you'll show up, adapt to the people who are already there, and fix the thing in front of you, there is room at Commure for you.

He joined Commure to work on a hardware deployment for a different team, and for his first stretch the rest of the company was something he knew about but had never touched. That changed on a project with a health system in New Orleans. The security staff there were required to write reports on every incident they responded to, and for a lot of them it was faster to say what happened than to write it down. Justin's manager told him to get in touch with Max Krueger, who leads Commure's forward-deployed engineering team, and see if Max could set up an Ambient AI account so the security team could dictate their reports instead. Fifteen minutes later, he had one. Not long after, Justin was sitting at the security chief's desk, plugged into his email, troubleshooting a USB microphone so the demo would run. He points to that moment as the day he actually joined Commure. When his original team's work wound down, he asked if there was any way he could work for Max, and Commure said yes.

Sales engineering at Commure is a wide-ranging job, which is part of why Justin likes it. He builds mockups and proofs of concept for demos, scopes integrations on the technical side of deals, configures accounts and templates, and helps with go-to-market work as it comes up. He works directly for Paul Pettit, who, on Justin's first day, walked him through a folder of internal tools and typed out prompts in real time to show how he thought about them. Justin describes the experience as catching fire. The team moves at Paul's rapid pace, and Justin found that not only could he keep up, he also loved the speed at which the team operated.

A less visible part of his week is release notes. Justin works with the product manager for Commure Pro to coordinate releases for one of the largest health systems in the country, which has strict validation requirements and an equally strict expectation for how release notes arrive.

Consolidating everything that changes in a release into one place, so that every rep and customer success manager is working from the same document, is work that seems small but prevents large problems. Some of that documentation was migrated from PatientKeeper repositories that are more than twenty years old. Asked which day at Commure he'll still be talking about in five years, Justin skipped the once-in-a-lifetime trip to Dubai and picked the release notes, because at the pace Commure's products change, the unglamorous documentation automation is the thing most likely to still be around.

The way he treats that work traces back to plumbing. Beyond the co-op, he spent weekends and summers doing residential repairs with his shop teacher, repairing fixtures and replacing heating systems in other people's homes. What stuck with him was the rule about the blue booties, which you put on in every house, no matter what you think of how the person lives. "It's always important to have that level of respect," he said. "You're a visitor on their turf." He feels the same thing now when he walks into a provider's office with two minutes between appointments to set up a macro, get them logged in, or take down feedback for the quality team. The people in that room have a reason to be there, and he is the one who has to adapt.

What has kept him at Commure is permission to be half-right early. He does his best work when he can float an idea before it's bulletproof, get a quick response, reach alignment, then start building. Rather than sitting around wondering whether there's a more polished way to say it. Commure rewards the authentic, timely answer over the perfect one. The expectation isn't that you never fail; it's that you never fail the same way twice.

He sees the same standard at every level of the company, from the CEO's late-night emails and Max's faithful weekly status updates down to the teams he's only ever worked with once or twice. He didn't always have it. It came from watching a nurse take a device he had just handed her and say she felt safer, and later from watching providers react to Ambient AI the same way.

"Don't get caught up on what your current title is; think about what your title could be in six to twelve months. That's how I would describe working at Commure."

‍

He and his fiancée bought their first house together around the same time he joined the sales engineering team, which replaced some of his old hobbies like tennis and hiking with new ones like gardening, shoveling, and installing dishwashers. He also found time in Dubai for an evening in the dunes with two colleagues, taking turns driving a rented off-road vehicle and paying a man with a falcon to let them pose for photos. He would do the dunes again. He would also, without hesitation, do the late-night hotel room troubleshooting again.

‍
Interested in a career building the next generation of healthcare technology powered by AI? We are always looking for talented people across our departments.

Explore Career Opportunities

Reid Gajewski was onsite at a hospital when he realized the magnitude of his role at Commure. He had just handed Commure's Ambient AI tool to a cardiologist with the busiest schedule he'd ever seen, a doctor who was literally running from room to room. The provider started a recording, walked into a patient's room, stopped it on the way out, and a few minutes later looked at the finished note. He told Reid it was perfect, and said he’d actually be able to spend time with his kids now. Then he turned to the health system executive standing next to them and said, "You better not take this away from us. Ever."

Reid is a Forward Deployed Engineer at Commure, a little over a year into his first job out of college, and moments like that are why he loves what he does.

His route here took some turns. In middle school, he wanted to be a politician. By high school, he was producing electronic music, buying synthesizers, learning how records were made, and eventually ghost-producing tracks for a few larger artists. Then he started selling products online and fell in love with digital marketing. The strategy behind positioning a product, finding the right audience, and making every dollar of ad spend work harder was fascinating to him. Still, the old political ambition carried him into college, where he studied political science and international relations, until he decided he'd rather build things than debate them.

The turn that stuck came in the spring of 2022, when he had an idea for a software product. He knew it could go viral, and he knew exactly how to market it. He just had no idea how to build it. So he taught himself to code, and through that process was pulled into the world of startups. Through college, he built a handful of small internet businesses; what he loved most was the full arc of deciding what was truly essential for a first version, getting it in front of people fast, and iterating until someone became the first customer. The job titles he imagined kept changing, but the theme never did. He wanted to build things from scratch and turn ideas into something people love.

"I was looking for a role where I could build, strategize, and own an outcome from beginning to end."

That theme is what made a conventional path feel wrong. Coming out of college, he was grinding coding interview prep for standard software engineering roles, and the more he thought about it, the less he wanted to spend his early career working on a small piece of a much larger system. He wanted to understand why something should be built, not just how. He wanted to sit with the people experiencing the problem, be held accountable for whether his solution worked, and grow beyond engineering into earning customer trust, communicating with executives, and making calls when information was incomplete.

That is essentially the job description of a Forward Deployed Engineer at Commure. FDEs build the product, then go on site with the clinicians and administrators who use it, own the outcome end to end, and act as their own product managers, shaping product direction rather than just executing on it. In a single day, Reid might work through complex billing scenarios with a coding team, brief a CFO on the revenue impact of reducing claim denials, and speak with Commure's CEO about where to attack next. After one conversation with Max Krueger, who leads the team, Reid knew immediately this was where he should be.

"There's almost never a detailed plan or a clear set of instructions, but there are hard problems to solve, ambitious goals to reach, and the trust and resources you need to make an impact."

Early on, Reid found himself in a frustrating loop with a customer: make the changes, get enthusiastic approval, then receive a new set of requests that seemed to contradict what they’d just approved. After several rounds, he stepped back to examine how clinicians beyond the pilot group actually worked. The problem became clear: he’d been tailoring the product to preferences that weren’t representative of the broader customer base. He brought the evidence to the customer, changed the rollout strategy, and saw adoption take off. Reid discovered that being responsive isn’t enough if you’re solving the wrong problem. Sometimes progress means questioning the feedback loop itself.

He thinks about the discipline required for the role, too. At a dinner with forward-deployed engineers from other companies, one veteran of the field described building an FDE organization that amounted to one-off modules for individual customers, none of it generalizable, none of it scalable. Reid sees that as the trap the role has to avoid. Commure works against this by identifying what each customer wants, triangulating it across other pilots, and turning it into something repeatable that can simply be switched on for the next customer who needs it.

The part of the job Reid says he loves the most is the payoff of all that work. FDEs build up deep, specific knowledge of whatever problem they're working on, then they get to walk into a room of doctors and healthcare executives as genuine domain experts and watch those people understand the value of what's been built. There's a very short distance between seeing a problem, deciding what should exist, and putting a working solution in front of the customer, something Reid says you rarely find at any company, and almost never in healthcare. Whether that's hours back in a physician's day or real money back in a health system's business, overseeing the implementation and the impact is what he joined Commure for.

The FDE role is not an easy one, so it helps that the Bay Area gives him somewhere to reset. You'll often find Reid on the Stanford Loop Trail at 7 am before work, driving through the mountains on weekends, or on skis, which he's been on longer than he can remember. 

More than a year in, Reid says the role has exceeded his expectations coming in. Commure handed him real accountability, ambiguous problems, and the freedom to find his own path forward, the exact environment he was looking for to start his career. Again and again, he stepped into situations he didn’t feel ready for, building confidence each time until he found himself operating in rooms he never imagined he’d reach this early in his career. It's the same environment waiting for the next person who'd rather own an outcome than execute a spec.

‍

Interested in a career building the next generation of healthcare technology powered by AI? We are always looking for talented people across our departments.

Explore Career Opportunities

In a January 2026 MGMA Stat poll, 48% of medical group leaders named denials and appeals as their biggest revenue cycle leak, and another 23% pointed to front-end problems like eligibility and registration errors. Billing and collections, the work medical billing software was originally built to do, came in at just 14%.

That distribution reads like a product roadmap. Legacy billing systems concentrate on the stage of the cycle that leaks least, while the expensive failures happen upstream, before a claim ever exists. AI-powered medical billing software closes those upstream leaks, and it's changing what health systems should expect from the entire category.

What medical billing software does (and where it stops)

Medical billing software converts clinical encounters into paid claims. It captures charges, codes services, checks claims against payer rules, submits them through a clearinghouse, posts remittances, and manages what patients owe.

Health systems typically run two billing streams through medical billing software at once. Professional billing covers clinician services on the CMS-1500 claim form, and institutional billing covers facility services on the UB-04, each with its own payer rules and edits.

Traditional platforms handle all of this as rule engines. They validate what staff enter and submit what passes validation, which works until the input is wrong. A claim built on an expired authorization, a missed charge, or thin documentation clears every edit and comes back weeks later as a denial someone has to work by hand. Denial management then becomes its own department, cleaning up failures the software waved through.

Payers spent the last several years automating their side of the table. Rule-based systems on the provider side can't keep pace with edits that change faster than staff can learn them, and that asymmetry is what the AI generation of medical billing software was built to correct.

What AI changes in the billing workflow

AI moves billing software from processing claims to preventing failures.

Before the visit, AI agents verify eligibility, confirm coverage details, and check authorization requirements against current payer policy. Those front-end errors ranked second among revenue leaks in MGMA's poll, and they're the cheapest failures to prevent because nothing has been billed yet.

Between the encounter and the claim, documentation quality decides what medical billing and coding software has to work with. Platforms like Commure Pro connect ambient documentation and charge capture at the point of care, so the billing record is built during the visit and coding starts from a complete note.

Before submission, AI compares the claim against the clinical record to catch charge capture gaps, then scores the claim against payer behavior to predict which claims will be denied while there's still time to fix them. A rules engine checks whether a claim is well formed, and a model trained on payer behavior checks whether it will actually get paid.

After submission, AI posts remittances, drafts appeals with payer-specific language, and ranks work queues by dollars at risk so your team spends its hours where the money is.

Some of this runs as a copilot, with a person approving each action, while the routine, high-volume tasks can run autonomously. Commure RCM applies that model as end-to-end medical billing automation, from eligibility through payment posting, with AI working each stage of the cycle in one system.

Medical billing software vs. RCM software

Medical billing software handles the middle of the revenue cycle. It creates claims, submits them, and posts payments. Revenue cycle management (RCM) software covers the whole arc, adding front-end functions like scheduling, eligibility, and prior authorization along with back-end denial management and analytics.

The distinction used to drive purchasing decisions, since practices bought billing tools while health systems assembled RCM stacks from point solutions. AI is collapsing it. A platform that predicts denials needs front-end eligibility data, and a model that drafts appeals needs the clinical documentation behind the claim, so AI-era platforms work the entire cycle as one system. When vendors use the two terms today, read them as a signal of scope, and evaluate any platform against the full cycle it claims to cover.

How to evaluate AI medical billing software

Searching for the best medical billing software mostly returns lists built for small practices (not enterprise health systems), and nearly every vendor on those lists now claims to be utilizing AI. What separates platforms that move revenue metrics from a renamed rules engine comes down to 5 questions.

Does the AI act, or only flag?

Dashboards that surface problems still leave the work to your team. Ask what the system completes on its own, from correcting a charge to filing an appeal, and ask which metrics moved for existing customers. First-pass yield, denial rate, and cost to collect are the numbers that prove a platform protects revenue integrity upstream instead of reporting on losses downstream.

What does it cost all-in, per claim?

Subscription price is a fraction of the real number. Per-claim fees, clearinghouse fees, implementation, and the staff hours the system still requires all belong in the comparison, because administrative overhead is where billing gets expensive. In a July 2026 Health Affairs article, researchers put US administrative spending at 25% of all healthcare costs, with billing for a simple office visit costing more than $20 per bill.

How deep is the EHR integration?

An AI platform that works inside clinical workflows can fix problems at the source, while one that ingests exports can only react to them. Also, be sure to ask how the platform behaves across multiple EHRs, because most enterprises run more than one.

Can the vendor show named results?

Marketing claims are cheap; reference customers aren't. One New York City health system lifted revenue by 20% and cut timely-filing denials by 53% with Commure RCM's charge reconciliation. Whatever platform you evaluate, ask for real-world results and the baseline they started from.

Where does the AI run, and what does it learn from?

Billing data is protected health information, so the model architecture is a compliance concern. Ask where models are hosted, whether your data trains anyone else's models, how outputs are audited, and how the vendor documents HIPAA compliance for its AI components specifically.

Medical billing software at health system scale

A system processing millions of claims a year turns a 1-point change in denial rate into millions of dollars, in either direction, so the gap between flagging and acting compounds with every claim.

Scale also multiplies complexity. Billing software for healthcare enterprises has to absorb multiple facilities, multiple EHRs, and both professional and institutional billing without fragmenting into siloed workflows. That's why billing is usually the highest-yield place to start a broader healthcare automation program, since the return shows up directly in cash.

Judge platforms on outcomes

Feature checklists made sense when every billing platform did roughly the same things. AI broke that equivalence, and the way to compare platforms now is by the revenue outcomes they can prove, measured in first-pass yield, denial rate, and cost to collect.

Hold every vendor to that standard, including us. See how Commure RCM automates the revenue cycle from eligibility through payment posting.

Discover Commure RCM
‍

Frequently asked questions

What is AI medical billing software?

AI medical billing software uses machine learning and AI agents to automate the work of getting claims paid, including eligibility verification, coding, charge reconciliation, denial prediction, and appeals. Traditional billing software validates and submits claims that staff must still prepare. AI platforms prepare, correct, and defend claims themselves, with people reviewing the cases that need judgment.

How long does AI medical billing software take to implement?

Implementation depends on EHR integration depth, claim volume, and how many facilities are migrating, and enterprise deployments typically run from a few weeks to several months. Ask vendors for a phased plan that starts with high-volume, routine claim types, since early autonomous wins build the case for the longer integrations.

Can AI fully automate medical billing?

Not yet, and vendors claiming otherwise deserve scrutiny. AI reliably automates routine, high-volume work like eligibility checks, coding for common encounter types, charge reconciliation, and payment posting. Complex cases, clinical judgment calls, and payer disputes still need people. The practical model runs routine claims autonomously and routes exceptions to human review queues.

How much does AI medical billing software cost?

Enterprise platforms price on claim volume, modules, and integration scope, so quotes are custom. Compare total cost across subscription, per-claim and clearinghouse fees, implementation, and the staff hours the system still requires, then weigh that against measurable movement in denial rate, first-pass yield, and cost to collect.

In CMS's April 2026 HCAHPS public report, covering nearly 4,000 hospitals, only 71% of patients said they would definitely recommend the hospital that treated them. Nearly 3 in 10 finished an inpatient stay not fully willing to recommend it.

Patient experience in healthcare is the accumulated result of every interaction a person has with your organization, from the first scheduling call to the last bill. These scores tell us those interactions still fall short.

This guide covers what patient experience is, how it differs from patient satisfaction, how it's measured, where it breaks down across the care journey, and what actually improves it.

What is patient experience in healthcare?

Patient experience in healthcare is the sum of every interaction a patient has with a health system across the full course of their care. It spans finding a provider, scheduling, intake, the visit itself, discharge, follow-up, and billing, and it includes every person and process the patient touches along the way.

Every interaction counts. Patients don't separate clinical care from everything wrapped around it. The hold time, the parking, the intake clipboard, and the clarity of the bill all land on the same ledger as the physician conversation.

Perception is the unit of measurement. Patient experience lives in what the patient noticed, and what your organization intended matters only as far as it showed. A discharge plan the care team considers thorough still fails if the patient gets home confused.

It runs the full course of care. Experience starts when someone first searches for help and continues long after the visit ends. Health systems that trace it stage by stage tend to find their biggest gaps between encounters rather than during them, which is why experience work overlaps so heavily with the patient journey in healthcare.

Patient experience vs patient satisfaction

Patient experience and patient satisfaction measure different things.

Patient experience asks whether something happened, and how often. Did your doctor explain things in a way you could understand? Did staff respond when you pressed the call button? These are observable events, which makes experience data comparable across patients and hospitals.

Patient satisfaction asks whether care met the patient's expectations. Two patients can receive identical care and report different levels of satisfaction because they walked in expecting different things. A procedure that felt rushed to one patient may have felt efficient to another.

Experience data tells you what to fix, while satisfaction data tells you how people felt about it. You need both, but improvement work runs on experience data because it points to specific, repeatable behaviors.

Patient experience is also distinct from the patient journey. The journey is the sequence of stages a patient moves through; experience is how those stages land. You improve the experience by fixing the journey stage where it breaks.

Why patient experience matters

Patient experience predicts outcomes, drives revenue, and shapes loyalty.

Start with what patients themselves say. In the Beryl Institute-Ipsos PX Pulse survey published in February 2025, 92% of U.S. consumers said patient experience is important to them, and most pointed to their own health and wellbeing as the reason. Patients treat experience as part of care quality.

The clinical evidence backs them up. A 2026 review in the Journal of Brown Hospital Medicine found that patient experience influences adherence to treatment, trust in providers, and clinical outcomes, with communication quality standing out as the single most influential factor across the hospital stay.

There's a hard financial link too. Since 2012, CMS has tied a portion of hospital reimbursement to HCAHPS performance through the Hospital Value-Based Purchasing program.

And when experience fails, patients act on it. In KFF polling from May 2025, 36% of adults said they skipped or postponed needed care in the past 12 months because of cost, and 18% said their health got worse as a result. Cost is an experience problem as much as a price problem. Confusing bills, surprise charges, and the absence of upfront estimates all push patients toward delay. Delayed care produces sicker patients and lost volume, which means a poor experience compounds into both a clinical and a financial liability.

How to measure patient experience

Patient experience is measured through standardized surveys, and in U.S. hospitals the standard is HCAHPS, the CMS survey that asks discharged patients about specific, observable aspects of their stay.

HCAHPS covers communication with nurses and doctors, responsiveness of staff, communication about medicines, cleanliness and restfulness of the environment, discharge information, care coordination, an overall rating, and whether the patient would recommend the hospital. Results are publicly reported on Care Compare, and because every participating hospital asks the same questions the same way, the scores support real comparison.

HCAHPS has a blind spot in that it only measures the inpatient stay, and it reaches patients days or weeks after discharge, relying on recall and fixed response options. Meanwhile, most of a patient's interactions now happen before and after the hospital, in scheduling, intake, messaging, telehealth, and billing, none of which HCAHPS asks about.

That's why measurement is expanding beyond the annual survey mindset. Real-time feedback after visits, analysis of patient messages and calls, and structured listening programs catch problems while they're still fixable for the patient who reported them. Treat HCAHPS as the benchmark and floor, and build faster feedback loops on top of it.

Where patient experience breaks down across the journey

Patient experience fails at predictable points, and most of them sit between clinical encounters rather than inside them.

Access and scheduling. The experience starts before anyone provides care, with whether patients can find you, book you, and reach you. Long hold times and phone-only scheduling lose patients at the front door; health systems are countering with self-scheduling and AI agents that expand call center capacity and patient access. This front layer of tools is what health systems are increasingly calling the digital front door in healthcare.

Intake and administrative work. Patients repeat the same information on paper forms, then wait while staff key it in. Administrative friction reads as disrespect for the patient's time, and it's one of the most fixable stages. Modern patient intake software moves registration, history, and consent to the patient's phone before arrival.

Communication before and between visits. One-way appointment reminders leave patient questions unanswered, and unanswered questions become no-shows and unprepared arrivals. Two-way patient communication software closes that loop, and AI-powered care journeys that guide preparation and answer questions have been shown to reduce no-shows while improving how prepared patients feel.

The visit itself. Communication dominates here. The 2026 Journal of Brown Hospital Medicine review found that empathy, active listening, and structured bedside communication consistently improve how patients perceive both the competence and compassion of their care team.

Discharge and follow-up. A rushed discharge can undo an excellent stay. Patients need clear instructions, medication reconciliation, and a follow-up plan, and the handoff between settings is where information gets dropped; well-designed transitions of care close those gaps. Post-discharge outreach matters more than most systems assume, since patients forget much of what they were told and questions inevitably surface at home.

Billing. The bill is usually the last interaction, so it colors the memory of everything before it. Upfront estimates, plain-language statements, and a human to call about a charge all belong in the experience.

How to improve patient experience

Improving patient experience means fixing the journey stage where your data says patients struggle, then verifying the fix moved the numbers. Programs that start from a specific gap outperform broad culture campaigns.‍

1. ‍Find your lowest-scoring touchpoint. Pull your HCAHPS results and whatever operational data you have on hold times, no-show rates, intake duration, and portal usage. Rank the gaps by patient volume affected.
   ‍
2. ‍Fix the mechanics before the messaging. If patients can't get through on the phone, scripting a warmer greeting won't help. Capacity, self-service options, and response times come first.
   ‍
3. ‍Automate the routine so staff can handle the human. Reminders, prep instructions, intake forms, and common questions can run automatically. That returns staff time to the conversations that genuinely need a person.
   ‍
4. ‍Close the loop with patients. Ask for feedback close to the moment, act on it, and show patients what changed. Feedback that disappears into a dashboard teaches patients to stop giving it.
   ‍‍
5. Measure again. Rerun the same numbers you started with so you can prove the fix worked. Tying experience gains to a metric leadership already watches, like no-show rate, HCAHPS composite, or call abandonment, is what turns a one-time project into funded, ongoing work.

The evidence that this works is concrete. Yale New Haven Health used Commure Engage to send bilingual pre-appointment text outreach for breast imaging, with reminders, prep information, and wayfinding down to parking directions. No-show and same-day cancellation rates fell 54%, from 13% to 6%. Patients kept appointments because the experience finally gave them what they needed to show up ready.

Pick a stage, remove the friction, and let the results argue for the next stage.

Patient experience is the service

Health systems compete on clinical quality, but patients can rarely observe clinical quality directly. What they can observe is the experience. How easy you were to reach, how prepared they felt, how clearly people spoke to them, how respectfully their time was treated. These observables drive their ratings, their loyalty, and increasingly their choice of provider.

Commure Engage gives health systems the front-cycle infrastructure for that layer, including scheduling, intake, two-way communication, and care navigation all on one platform. This helps to ensure every stage of the journey works the way patients expect. If patient experience is on your roadmap this year, start where your patients already told you it hurts.
‍

Discover Commure Engage

‍

Frequently asked questions

What is patient experience in healthcare?

Patient experience in healthcare is the full range of interactions a patient has with the healthcare system. It measures whether specific aspects of care actually happened, such as clear communication and timely access, across the entire continuum of care.

What is the difference between patient experience and patient satisfaction?

Patient experience measures whether specific interactions happened and how often, such as whether providers explained things clearly. Patient satisfaction measures whether care met the patient's personal expectations. Two patients can receive identical care and report different satisfaction, which is why experience data drives improvement work while satisfaction data captures sentiment.

How is patient experience measured?

Patient experience is measured primarily through standardized surveys. In U.S. hospitals, the CMS HCAHPS survey asks discharged patients about communication, responsiveness, cleanliness, restfulness, medications, and discharge information, and results are publicly reported. Many organizations add real-time feedback tools, post-visit surveys, and analysis of patient messages to catch issues sooner than annual survey cycles allow.

Why is patient experience important?

Patient experience is linked to treatment adherence, trust, and clinical outcomes, and it carries direct financial weight. CMS ties a portion of hospital Medicare reimbursement to HCAHPS scores through Value-Based Purchasing. In a 2025 Beryl Institute-Ipsos survey, 92% of consumers said experience matters to them, citing their own health as the reason.

How can health systems improve patient experience?

Health systems can improve patient experience by finding the journey stage where their data shows patients struggling, fixing the underlying mechanics, and remeasuring. High-impact levers include self-scheduling, digital intake, two-way communication with reminders and prep guidance, structured discharge follow-up, and transparent billing.

For a decade, the best engineers in tech went to consumer, cloud, or autonomy. Now they're going into healthcare due to the sheer size of opportunity to make an impact. A growing share of them are landing at Commure.

We track every hire back to where they came from, and the map has gotten dense fast, with big tech, big healthcare, defense, finance, and semiconductors, all pointing the same direction.

In the last 12 months, we've seen a wave of incredible engineering talent join from a broad range of world-class companies: Meta, Google, Uber, Tesla, Apple, Carta, Epic, Verkada, Scale, Palantir, SpaceX, Stanford, MIT, Berkeley, and more.

Healthcare Is the Hardest, Highest-Leverage Problem Left

Engineers and top talent in any domain go where the leverage is highest. 

U.S. healthcare is a $4.5T industry that hasn't moved on productivity in two decades. Physicians work 57.8-hour weeks, and 7.3 of those hours go to prior auth, insurance forms, and admin. The 2025 CAQH Index put administrative savings from electronic transactions at $258B in 2024, with another $21B sitting on the table, not automated. Denials are climbing. Documentation eats a quarter of a clinician's day. Charge capture leakage costs the average health system millions a year in care delivered and never billed.

Every one of those numbers is a rate-limiting bottleneck software can remove. Healthcare is the last industry this size still waiting to be rebuilt, and Commure is building the AI-native platform to do it: ambient documentation, autonomous coding, AI call center agents, staff safety, and end-to-end RCM, running across 60+ EHR systems. One well-designed model saves an hour of a clinician's day. One well-run agent recovers millions in leakage. One well-instrumented workflow reaches 100 million patients. That's the pitch. It's landing.

Talent Cohorts

Talent moves in patterns. Here are a few we've noticed in the last few months:

Meta. Multiple engineering hires this year, most last quarter. They built ranking and inference at a trillion-event scale. Here they get the same rigor pointed at hours returned to a clinician's day instead of engagement metrics.

Uber. Wrote the playbook on real-time ops and marketplace dynamics inside a regulated industry. Now leading central ops for AI deployments across 130+ health systems.

Tesla / SpaceX. Full-stack thinking, vertical integration, comfort with the fact that the hard part is everything around the model.‍

Founders. A dozen-plus acqui-hired startups. They hit distribution walls solo; Commure gave them scale. Ex-founders bring some of the best talent in the company. 

Why Talent Stays

Scale. Tens of millions of U.S. healthcare appointments now run on Commure. HCA and Tenet, two of the largest for-profit systems in the country, partner with us. 100M+ patient interactions a year. 

Speed. Five acquisitions, the business has doubled growth every year for the last 3 years, and we still move with our “speed of all else” mantra. Ownership sits with individuals, not committees. Prototype to production inside a sprint. "I want to move faster" is the single most common reason big-tech hires give us.

Distribution. We're already inside 130+, with 60+ EHR integrations live and 350,000 clinicians on the platform. Shipping is only fun and impactful when someone uses what you built.

Stakes. $25B in annual claims processed. Every transaction is care delivered to a real patient. Get it right, and clinicians spend more time with patients, health systems stay funded, and doors stay open.

What's Next

Healthcare AI consolidates around a handful of companies with the platform breadth, distribution, and talent density to actually move care delivery. Commure intends to lead that consolidation.

We're hiring across engineering, product, ML, sales, customer success, and ops.
‍

See open roles →

When Cem Torun started on what would become Commure's Ambient AI product, he spent the first 4 months tasked with figuring out whether the idea was worth building at all. There were no users and no team, just Cem and a rough concept that Tanay, Commure's CEO, and Hersh, its CBO, had pointed him toward before stepping back to let him run with it.

Four years later, he's the VP of Engineering for Ambient AI and leads a team of more than 40.

When he first moved to the Bay Area after college, he skipped the apartment search and lived on a retired pharmacist's boat in the Redwood City harbor for 2 weeks, commuting to the office by Uber and turning up most mornings covered in mosquito bites. The owner had spent years in pharma, so their evening conversations about the industry ended up teaching Cem a fair amount about the field he'd just joined.

His route into healthcare took a few turns along the way. As a kid, he wanted to be a YouTuber, talking his teachers into letting him turn writing assignments into short films instead of essays. His parents wanted a doctor in the family, so he spent his first 2 years of college in pre-med, building on the time he'd spent volunteering at a palliative oncology center in high school.

That volunteer work showed him how many problems healthcare had, but fixing them one patient at a time felt too small for the scale he wanted, so he taught himself to code and figured out he could build technology for medicine instead of practicing it.

Being able to work on the first iteration of Commure’s Ambient AI kept rewarding his willingness to take on new problems and think outside of the box. After 4 months of iterating, he had proven the idea was worth pursuing and was told to build a team around it. When that team grew to 6 engineers and still had no product manager, he covered the role himself until the company hired one. He did the same with design, building it to a solid standard until a full-time designer came on board.

‍

Cem had his biggest test at Commure with the first major Ambient AI enterprise deal. Getting a large health system live was a big implementation, so Cem flew out and worked alongside the team in person for 4 to 6 weeks straight to make sure everything was specifically tailored for their unique workflows.

When Cem talks about why he's stayed 4 years, he says it's uncommon for someone early in their career to be handed stakes this high, and he's watched Commure choose to grow its own leaders instead of bringing in a 20-year veteran every time a hard problem surfaces.

He tells candidates their impact won't be capped by title or tenure, and that no one waits around for an assignment. If you see a problem worth solving, you go solve it. He's honest about the trade-off too, that people who measure their work in weeks tend to miss the results, while those who stay on a hard problem for a couple of years see what that kind of focus produces.

He points to his work on Ambient AI as the clearest example of this. The product supports some of the country's leading oncology and drug-discovery organizations, so the code his team writes reaches cancer research and patient care directly. He admits he undervalued that impact early on, since it was his first full-time job and he assumed most companies operated at those high stakes.

Cem now runs his own team on the same logic, giving people work that sits a little past what they've done before and puts enough structure around it so no one falls too far. As they show they can move quickly, he hands them more.

That approach is how a solo, 4-month experiment turned into a 40-person engineering org in less than 3 years, and how the person who started it went from proving out an idea by himself to leading the team that scaled it.

The same drive holds outside the office, where Cem is training for his first triathlon. He took it on because he wanted something genuinely hard, then talked 6 friends into doing it with him. He booked the race in Berlin because a domestic one sounded boring.

Nine months ago he couldn't swim, and the first time he got in a pool he couldn't make it a single length without stopping for air, which he says made him feel like a kid again. Since then he's learned to swim, bought his first real road bike, and picked up enough about gear sets, shifters, and fixing a flat to give a TED talk on it.

It's the same move he's made his whole career, going after the thing he doesn't know how to do yet and working at it to make it happen.

‍

Interested in a career building the next generation of healthcare technology powered by AI? We are always looking for talented people across our departments.

Explore Career Opportunities