eClinicalWorks projects 30% efficiency gains with AI, Canvas upgrades

Healthcare IT Specialist Demonstrates Electronic Records To Diverse Medical Team.

...

Alexis Kayser

By Alexis Kayser

Healthcare Editor

See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.

This is a preview of the October 29 edition of Access Health—Tap here to get this newsletter delivered straight to your inbox on Thursdays.

Good morning. Last week, I traveled to Nashville for eClinicalWorks' National Conference, where the EHR vendor laid out an ambitious vision for the future of health care technology: fewer clicks, more automation and AI that even the smallest medical practices can afford.

In Other News

Major health care headlines from the week

The Department of Health and Human Services finalized a rule Monday aimed at reducing regulatory burdens on health IT developers and encouraging innovation. The Health Data, Technology, and Interoperability (HTI-5) rule is projected to save more than $1.5 billion in regulatory costs, according to the agency.
- The rule eases administrative and compliance requirements for developers participating in the federal Health IT Certification Program. HHS says the changes will allow companies to redirect resources toward improving technology for clinicians and patients, including tools that support interoperability and automation.
- More changes are on the horizon. The Office of the National Coordinator for Health Information Technology plans to propose additional measures to improve electronic health information sharing, expand the use of application programming interfaces (APIs) and update certification requirements.
- However, information blocking remains unfinished business. HHS said it is still reviewing public comments on information-blocking proposals included in the original HTI-5 proposed rule, signaling that additional regulatory action could follow.
- You can view the final rule in the Federal Register, and this webpage provides additional context.
A liability battle is emerging between hospitals and AI developers over who is responsible when autonomous AI tools cause patient harm. Dr. Brian Anderson, CEO of the Coalition for Health AI, said at Newsweek’s September AI Health Summit that health systems and leading AI developers are already at odds over control and accountability. "There's a standoff right now going on with a lot of health systems and the frontier labs about who's going to be responsible and control the agents," he said.
- Under current legal frameworks, health care providers generally remain liable when something goes wrong with an AI tool, Anderson said**. But increasingly autonomous agents complicate that arrangement, particularly when clinicians cannot explain an AI system's decisions.** The American Medical Association has argued that accountability should fall on the party best positioned to understand and mitigate a technology's risks, rather than automatically resting with physicians.
- Shared liability could become more common as AI takes on higher-risk clinical tasks. Anderson pointed to Utah, where officials are considering shared responsibility and companies are testing more autonomous applications. Doctronic, which is testing AI-assisted prescription renewals in the state, has obtained malpractice insurance, a development Anderson suggested could signal growing recognition among technology companies that they may need to assume some liability for their products.
- Get the full scoop at Newsweek.
Dr. Jennifer Shuford, deputy director and chief medical officer of the Centers for Disease Control and Prevention, is resigning after just six months on the job, according to an internal agency email obtained by The Guardian. Shuford, appointed by President Donald Trump in April, is expected to depart Nov. 12, citing a desire to spend more time with her family.
- Her departure marks another leadership shake-up at the CDC, which has faced significant turmoil under Health Secretary Robert F. Kennedy Jr. Shuford was appointed as part of a four-person leadership team intended to stabilize the agency following the departure of former Director Susan Monarez, who publicly criticized Kennedy's directives as "unscientific."
- The resignation comes amid sweeping changes to the nation's public health agency, including the replacement of all 17 members of its vaccine advisory committee and the loss of approximately one-third of its workforce. Dr. Satish Pillai will serve as acting deputy director and chief medical officer while the agency searches for Shuford's successor.
Flu season is off to an unusually early start in the U.S., raising concerns about hospital capacity heading into winter. The CDC reported 4,811 laboratory-confirmed influenza hospital admissions during the week ending Oct. 17, with hospitalization rates reaching their highest level for this point in the season since 2010-11. Walgreens separately reported that flu activity is tracking roughly four weeks ahead of recent seasons.
- H1N1 is the dominant influenza strain circulating so far this season, accounting for 239 of 241 influenza A samples analyzed by the CDC during the week ending Oct. 17. Health officials say it is too early to determine whether the season will be more severe than recent years or how genetic changes in the virus might affect vaccine effectiveness.
- Hospitals should prepare for increased emergency department visits, longer stays and potential disruptions to elective procedures, Dr. Bruce Farber, chief public health and epidemiology officer of Northwell Health told me on Wednesday. He recommended monitoring community test positivity rates to anticipate surges in patient volume.
- Farber will be back for a Pulse Check in next week’s edition—stay tuned, and send questions about this year’s flu season to a.kayser@newsweek.com.

Pulse Check

Executive perspectives on key industry issues

Girish Navani is cofounder and CEO of eClinicalWorks.

...

I sat down with Girish Navani, CEO and cofounder of eClinicalWorks, during the company's National Conference on October 23. We discussed the company's expanding AI ambitions, its push to make advanced technology accessible to smaller, independent medical practices and how Navani is thinking about the future of health care after nearly three decades at the helm.

Read more about the company’s game plan in this article, and read on for a transcript of my interview with Navani.

Editor's Note: Responses have been lightly edited for length and clarity.

Of all the new features and business units announced today, which are you personally most excited about?

For 27 years, we've built electronic health records with what you would call the traditional enterprise software mindset, and what we showcased today is not enterprise software the way you and I know it.

I think you can put it into analogies like that's the iPhone moment of a smartphone business, or another very, very significant jump in a technological curve.

To me, 12.04's Canvas design is groundbreaking. It will probably lead to others doing something similar over time, but it changes the whole user experience where you don't have to spend as much time on your computer.

That [computer time] is the number one problem in health care, so addressing it in a systemic way with AI and Canvas design that just takes away the whole concept of "click, click, click" is the big [announcement].

Obviously, being able to write your own agents [in AI Workbench]...it's going to be mesmerizing to see what customers come up with. But internally, we've got a list of 60 [agents] already created within the last two and a half weeks.

So the whole idea that you can build software in a way that has agent-level capability to reason, think, do, plan and act, be able to do that within an hour or two, and just make it run in a scalable environment, in a secure environment, with the guardrails that we provide, with a click and say "publish," that is something I haven't seen somebody else do yet—in health care, at least.

Both changes come with some excitement, but they also come with some level of planning that needs to go into the latter part of this year. How do we scale this knowledge and education to 17,800 customers?

You mentioned that you expect customers to achieve a 30-percent improvement in practice efficiency after upgrading to 12.04. How soon after the upgrade do you expect them to see those gains?

One quarter.

We've seen it internally, so I speak as an optimist, but a realist, too. We have done similar changes to our own internal organization, though we're not a health care delivery organization.

Given what we do internally with technology transfers and changes, from software development to support, I think that if customers take [the upgrade] in Q1, they should see a visible difference by the end of Q2.

How they take that savings and translate it to growing, adding more doctors, adding more capability, seeing more patients or finding other things to do with it, is a choice that the health care institution makes.

But the number 30 [percent], while it sounds high, is not unrealistic.

We've done some internal benchmarks on just the revenue cycle management side. The early number is 50 [percent]. I could do double the work with the same number of people.

That is not an insignificant game changer.

So, to me, the whole idea of agents and agentic workflows being deployed in the organization creates a radical shift in efficiency. Health care needs it. It can be done.

I don't know of anybody with a product like Workbench working with a broad ecosystem, not just [their own clients]. We're not writing agents that only do work for the eClinicalWorks EMR users. This connects into Salesforce, Microsoft 365, another EHR you may have. Or it may be you're part of an ACO, and this is many different electronic health records.

The fact that you can do value-based care with an agent across practices, across EHR systems, is just a revolutionary product line. And I think we're ahead because, as I follow along the industry, I don't know of anybody who's got the breadth of connectors we already have live, or the breadth in capability we're supporting today, or the whole agentic ecosystem that comes with it, the cloud.

It's the whole stack. It's not just the software.

How many of your customers do you expect to upgrade to 12.04 following its initial release?

The majority will upgrade in the first year. Typically, in the first six months, you will see 60, 70, 80 percent of them get there. That's the speed at which we do it.

For example, [92 to 93] percent of my customers are on 12.03, which was released last year.

The speed of release is only predicated by the customer's appetite to absorb it, not by the limiting factor of when we give it.

We start Jan. 15, so the first two weeks are usually a little bit more cautious, and then it just shoots up. It shoots up to like 500, 600 [upgrades] being done in a day, and it's a press of a button eventually.

It's when customers say they're willing to take it, because we give them all the education material ahead of time. The Smart University [eClinicalWorks’ official online learning platform for clients and partners] gets upgraded, so people know what they're getting, why they're getting it and how they're going to use it.

We expect we'll do the majority of our upgrades in the first half of the year.

What enables that level of efficiency at eClinicalWorks? How do you keep costs low and encourage customers to upgrade quickly?

You earn their trust. This is not something you can just get. It's not something that is handed down to you because you're special. It's something you earn over a decade or two decades of doing work.

Our customers today just trust the fact that when we give them new capabilities, we'll make it better for them.

Now, this is a giant leap, so we're going to tell people to make sure that they prepare themselves to use these capabilities. But the trust is there.

If you told us to upgrade everybody in the first two weeks, we could do it, but that's not the speed customers will go at.

And how do you keep costs low?

I think that's a fascinating question, because we do keep our costs low, even with AI. We sell AI at a much lower price point with all of our products.

In fact, you saw that 12.04 is going to include Smart Canvas and Form Builder and many other AI capabilities at no additional cost. That's like eClinicalWorks absorbing the whole cost structure there.

So how do we do it? First, we're growing at about 12 percent. Any business that grows can afford to do what we are doing.

Second, we are a private company with no shareholders, so I have two choices. I don't have three. I don't have to worry about shareholders. I only have to worry about employees, and I have to worry about my customers.

As long as I am financially rewarding the employees at eClinicalWorks, and as long as I'm financially rewarding my customers, the company is in a good place.

eClinicalWorks will continue to do two things well: make sure its employees are financially rewarded, and make sure our customers can afford the products we build. It is true. It is not marketing. This is factual.

We call it “democratizing AI” now inside our company. We want to make AI affordable to even the solo practice. We don't want to make it hard. We just want to make it part of their system, and you will see us keep pushing that envelope.

The top-line growth helps, and the fact that we run a fiscally responsible company. And we now added new business lines. Pharmacy is going to be a big growth engine. Inpatient is going to be a big growth engine.

People sometimes ask me, "Oh, is inpatient too late?"

Four people signed up to join with me and go live in the next quarter. There are at least 25 others waiting that want to talk to us about doing it.

I've come to this realization over three decades of doing the work: If you build good products with good service at good price points, you can always create a market.

Why are those inpatient hospital customers interested in switching to eClinicalWorks?

That tier of the market has been ignored by the vendors. I don't want to get into specifics about the vendors, but I think it's a market that has been underserved with modern technology, and I think there are at least 2,000 hospitals that fall into that bucket.

People desperately want to figure out how to switch to a cloud-first, an AI-first product. They don't want the legacy. It costs too much. It's antiquated in design. They want to figure out something else. So that segment of the population of hospitals definitely needs [products like ours].

Time will tell. I'm not saying it's going to be easy for [eClinicalWorks] to keep iterating and making [customers] delighted all the time. Customers bring responsibilities. But having a start with four already wanting to go live on the early journey, and at least two dozen waiting to talk to us in the next quarter, that's significant.

It's the same thing on the pharmacy side. I just finished meeting some people, and they all said, "Why don't we just do a meeting in January? We'll bring a lot of our health center customers that need the pharmacy products."

So I would love to get 100 pharmacies deployed next year.

I often hear hospital executives argue that the industry needs to fix its underlying technology infrastructure rather than simply building AI on top of it. Do you agree?

One hundred percent.

Bolting on AI on top of a legacy architecture of technology is not going to work.

It will make some marginal improvement. Some people can be happy with that two-percent, three-percent, five-percent efficiency gain and go, "Wow, I attained it."

But we're talking 30. You're not going to attain that 30 percent by bolting on AI.

It has to be immersive. Your core day-to-day workflows have to be changed and adjusted with agentic processes. It cannot be any other way.

But it's going to require change. And then my human trepidation of change may kick in, going, "No, no, no, that's mine. I've got to do it this way."

That's the word: courage.

If your leader does not have the courage to tell its organization to change in this new paradigm of seismic shift in technology, that organization is going to fall behind, big time behind.

There's going to be transformations in this space.

Yes, it's a regulated industry, so it won't happen at the speed that BlackBerry and iPhone did, where one pretty much killed the other overnight. It will take time.

But there's a new set of groups and practices coming up that are putting native AI on everything they do.

They cannot take a legacy system of record and just bolt on three things and say, "Great, we've got AI built in."

So, talk systemic change. Look at key workflows that are administratively burdened. Look at revenue cycle management being done through a complete agentic process.

Transform your front office booking with voice agents and even the scheduling to be driven by AI agents finding the right slots.

And you may be delivering health care at 30-percent efficiency within a quarter, not even a year.

This is not a futuristic vision, "Come, let's create Disney magic." No, this is real. Let's go do it tomorrow, and you will see a perceived difference within 12 weeks.

How do you sense clinicians feel about allowing AI to take control of certain parts of their workflows?

There's always a bell curve.

I wouldn't say you could generalize this or simplify this with just clinicians. We're all human beings, so we all fall on a distribution curve of how we think.

Some will be early adopters, the majority will be middle adopters, and some will be laggards in implementing it.

Nothing's right or wrong about it. That's the way we are. That's the way I am. There are things I adopt very quickly, and there are things I wait for. That's just me.

So I'm not going to suggest that everyone's going to think the way I want them to think when it comes to the optimism I have, that including intelligence in your product and in your day-to-day is going to be transformative to your efficiency.

You need to make that decision, but I would ask the question differently: What else are you going to do if you don't do it?

Status quo?

Well, then you are waiting in the middle of the railroad track, expecting the train to come. It's just a matter of time.

How long do you want to stand on the tracks?

I was interested in some of the new features in healow. [Ed.’s note: During the keynote address, Navani announced that the patient portal for healow— eClinicalWorks’ patient-facing digital health platform—has been redesigned to create a more cohesive experience under each practice’s branding. He also said that the healow Genie AI assistant has expanded beyond AI-powered phone requests to support text, WhatsApp and email interactions with patients for routine requests.]

Do you anticipate a future where patients choose their providers based on the ease of the user interface or the digital tools available to them?

Yeah, one is ease of everything we talked about. Second, visibility of price.

There's just one thing missing, but I think that's going to change, because the [Trump] administration is pushing it: price.

Health care is the only [business] segment where a consumer spends money and they don't know [the full cost] upfront.

We need to build digital experiences for easy booking, easy check-in, easy payment. But we also need to make it easy for the patient to know what it's going to cost.

And the administration is pushing hard on price transparency, and I applaud them for doing it, because there is no reason why a consumer shouldn't know exactly what a colonoscopy is going to cost or what their MRI is going to cost.

And if you are going to charge more and somebody is going to charge less, the consumer should know it.

We need to add that to healow. We need to add that to the website of every office.

I'm optimistic it's going to happen, and then your personal agent will find you the cheapest MRI place to go to that's most available in a timely fashion.

So, along with the work we have done in creating a very holistic platform for booking, check-in, appointments and all, I think one more thing needs to be added.

But I think it's going to be added within the next quarter. I hope.

I'm hearing increasingly that that's an area where there is a lot of onus and priority being put in, but price transparency hopefully gets done.

And when it gets done, it really results in a consumer then deciding, “Where do I want to go get my care?”

Because price is going to be a component. Always is. Convenience is one, access is one, but so is price.

I noticed that nearly everything demonstrated onstage is expected to be available within the next quarter or in January. At other EHR conferences, some announcements concern products that won't be available for years. Is there a specific reason you wait to announce products until they're nearly ready?

At least for the last three or four years, that's been who we are.

There is no use in me telling you today on this keynote stage if something's coming out in 2028. Seriously, there isn't.

You've taken the time, your money and your three days to be with me. You want to go back and do something with it.

I think it's just philosophical. It doesn't add up in my mind that you should announce something two years down the road that you've not finished yet.

I'm not saying it's wrong or right. Let people decide. I'm just telling you the way we think.

eClinicalWorks sees that if you can release something in the next quarter, great. Then talk about it, showcase it, let them go back and do it.

When they come next October, we'll talk to them again about what they can do by January, and that's become more of a repeatable pattern within the company.

Typically, we announce it in October. Normal rollouts start by Jan. 15, and they're usually 90 percent done by the time the second quarter ends.

Look at [clinical ambient documentation tool] Sunoh's numbers: 67 percent of the people in the audience now use Sunoh.

It's just two years ago when we showed it live onstage. People were like, "You're building this? Oh, we buy third-party. When did you start working on it?"

Well, we started working on it a year ago. We just didn't want to tell you about it because it was not something we tell you until we finish it.

But when we finished it and we liked it, now we show it onstage, and then it just takes off.

Genie [AI assistant that handles routine patient calls and tasks like appointment scheduling] is just like that. Genie was in the works for almost two and a half years. Until we got it right, we just did not want to discuss it.

So yes, there's a temptation to go show off onstage, but there's enough to show that we didn't need to worry about it.

Sometimes we make mistakes. Don't get me wrong. We make a mistake or we announce something that we think is ready, and then when we go and do early deployments, we realize that we could have done better.

But that's rarer now than normal. That can still happen. It's a tech business.

The way they say in tech, if you hit one out of 10 shots, you're considered successful. We're talking about hitting nine out of 10 shots, so we're good.

You mentioned that several of these new AI capabilities will be free. Are they genuinely free, rather than subject to token-based charges?

Different products, different [pricing].

Workbench is not free. Workbench software, but you pay for tokens. I cannot eat their Claude tokens. I'll end up standing in line for debt financing like all the other companies are.

I can't give you Claude tokens for free. It's usage-based.

But Smart Canvas, the canvas design and all the interpretation we are doing, free.

It's going to cost a lot of money, but it's free.

[And] healow IQ, if you are licensing Sunoh, or if you license our AI bundle already, no additional cost. It's part of the two bundles.

The whole form designer, which can be prompted to build new forms, is using tokens. Free.

Smart scheduling, a whole scheduling interface with recommendations that are being calculated behind the scenes, are using tokens. Free.

But we calculated what we think the token utilization there would be, and we said we can afford to eat that [cost].

So it wasn't radically unthought-through or ill-planned.

We said that's a predictable token cost. How many telephone encounters does a doctor get? So, we can calculate it.

Workbench, we don't know. People will build agents that will just go do 500 things and all that is token-based. We cannot [predict those costs and absorb them].

I think there's some credits being given so that people can start experimenting with it, and then it's no different than anything else that you see in the market, whether it's OpenAI, Claude, Replit. Even Microsoft products that are now in the AI world, Copilot, it's based on what you use.

But you use less, you pay less.

So yeah, some of it is really, really free, and it is something we thought we could afford to do.

We're doing it at scale right now. We'll see where the world goes for us.

C-Suite Shuffles

Where health care leaders are coming and going

Syed Samiullah was named chief digital and information officer at Summa Health, based in Akron, Ohio.
- Samiullah joins Summa from Optum Health, where he was chief information officer for the East Region.
Ardent Health in Nashville tapped Gaurav Khanna as chief digital and information officer.
- Khanna most recently served as vice president and segment CIO for Humana’s Centerwell Pharmacy business.
Dr. Pravene Nath was appointed chief innovation and physician executive at Rady Children’s Health in San Diego and Orange, California.
- His resume includes senior leadership roles at Walmart Health, Roche, Mindstrong Health, NYU Langone Medical Center and Stanford Health Care, where—while serving as chief information and digital officer—he led an enterprise EHR transition and launched the MyHealth patient platform.

This is a preview of the October 29 edition of Access Health—Tap here to get this newsletter delivered straight to your inbox on Thursdays.

View 0 Comments

Request Reprint & Licensing

Submit Correction

View Editorial & AI Guidelines

Read More on Access Health Newsletter