OpenAI CEO Sam Altman speaks during a keynote address announcing ChatGPT integration for Bing at Microsoft in Redmond, Washington, Feb. 7, 2023.
Jason Redmond | AFP | Getty Images
OpenAI on Monday announced its biggest news since ChatGPT's debut: It's launching ChatGPT Enterprise, the AI chatbot's business tier, available starting Monday.
The tool has been in development for "under a year" and had the help of more than 20 companies of varying sizes and industries, OpenAI COO Brad Lightcap told CNBC. ChatGPT Enterprise includes access to GPT-4 with no usage caps, performance that's up to two times faster than previous versions, and API credits. Lightcap said that pricing would not be publicly announced and that it "it will depend, for us, on every company's use cases and size." Beta users included Block, Canva and The Estée Lauder Cos.
Earlier this year, Microsoft's expanded investment in OpenAI — an additional $10 billion — made it the biggest AI investment of the year, according to PitchBook, and in April, the startup reportedly closed a $300 million share sale at a valuation between $27 billion-$29 billion, with investments from firms such as Sequoia Capital and Andreessen Horowitz. Two months after ChatGPT's launch in November, it surpassed 100 million monthly active users, breaking records for the fastest-growing consumer application in history: "a phenomenal uptake – we've frankly never seen anything like it, and interest has grown ever since," Brian Burke, a research vice president at Gartner, told CNBC in May.
More than 80% of Fortune 500 companies had teams actively using ChatGPT, per Lightcap and OpenAI.
One key differentiator between ChatGPT Enterprise and the consumer-facing version: ChatGPT Enterprise will allow clients to input company data to train and customize ChatGPT for their own industries and use cases, although some of those features aren't yet available in Monday's debut. The company also plans to introduce another tier of usage, called ChatGPT Business, for smaller teams, but did not specify a timeline.
Lightcap told CNBC that rolling out the enterprise version first, and waiting on the business tier, "gives us a little bit more of a way to engage with teams in a hands-on way and understand what the deployment motion looks like before we fully open it up."
OpenAI noted in a blog post that "We do not train on your business data or conversations, and our models don't learn from your usage," adding that clients' conversation data would be encrypted both at transit and at rest. The company does, however, log aggregate data on how the tool is used, including performance metadata and more, as is relatively standard, Lightcap said.
ChatGPT Enterprise's debut comes as the AI arms race continues to heat up among chatbot leaders such as OpenAI, Microsoft, Google and Anthropic. In an effort to encourage consumers to adopt generative AI into their daily routines, tech giants are racing to launch not only new chatbot apps, but also new features. In May, OpenAI launched its iOS app, followed by the Android app in July. Google is regularly rolling out updates to its Bard chatbot, and Microsoft is doing the same with Bing, introducing features like visual search. Anthropic, the AI startup founded by ex-OpenAI executives, debuted a new AI chatbot, Claude 2, in July, months after raising $750 million over two financing rounds.
When asked how ChatGPT Enterprise compares with Bing Chat Enterprise, Microsoft's enterprise AI chatbot, an OpenAI representative told CNBC, "This is an OpenAI product independent of Microsoft. That said, we hope our product can work with other tools including Microsoft's. Customers can choose which platform is right for their business."
ChatGPT, like many large language models, is expensive to operate, with each chat likely costing OpenAI "single-digit cents," according to a December tweet by CEO Sam Altman, suggesting that operating the service for 100 million people a month could cost millions of dollars.
The biggest obstacle to ChatGPT Enterprise's development was figuring out how to prioritize features, Lightcap told CNBC.
Out of all the things shipping in the next couple of months, he said, "the prioritization of how you pulled forward those things based on how people are using the product — and what people really want and what's empowering — was the topic of a lot of debate, I would say, on the team."
One concrete example is Code Interpreter, a ChatGPT Plus feature that has since been renamed to Advanced Data Analysis. Lightcap said that the team questioned whether the feature was a priority for ChatGPT Enterprise and that it "sat stack-ranked in a list with a bunch of other things that we think are kind of equally or more exciting," but companies' feedback caused them to prioritize offering it sooner rather than later.
OpenAI plans to onboard "as many enterprises as we can over the next few weeks," per the company's blog post.
Virtual sitting and virtual nursing solutions are receiving attention from healthcare organizations. Although the use of virtual nursing especially is not yet widespread, and although both use cases often require a significant up-front investment in equipment, organizations are interested in their potential to improve patient outcomes, alleviate staffing shortages, and reduce costs.
This report—KLAS’ first about virtual sitting and nursing—is an early look at customer experiences and the positive outcomes (clinical, financial, and operational) organizations are seeing in this space. Broad market insights are also shared. KLAS will continue collecting data on virtual sitting and nursing software and share insights as the market grows.
Understanding and Analysing Virtual Sitting and Virtual Nursing Solutions: Key Insights
Longtime Vendor AvaSure Drives ROI via Patient & Staff Safety Outcomes, though Some Respondents Mention High Hardware Costs: AvaSure is a well-known player in this market, with customers using their solution for an average of five years. Most of these customers are large organizations utilizing the system for virtual monitoring. They are pleased with its ability to produce positive outcomes, such as reducing patient falls, preventing tube/line issues, deterring unauthorized drug use, and ensuring staff safety by curbing patient aggression. The system also offers financial benefits by enhancing the capacity to observe patients and prevent adverse events. Customers appreciate its insightful analytics, two-way patient communication, multilingual pre-recorded messages, and the convenience of a one-stop shop for equipment. However, some AvaSure users express concerns about the system’s cost, feeling that they are being charged extra for cameras and necessary equipment. Despite this, they plan to continue with AvaSure due to the return on investment and positive outcomes they have experienced. Their decision to stay hinges on manageable ongoing costs related to equipment maintenance, replacement, and organization-wide scalability. Integration with their Electronic Health Record (EHR) may also influence future decisions for some respondents. Feedback on support is mixed, with some noting slower and less proactive assistance as AvaSure has grown. Additionally, users find portable cameras cumbersome in smaller rooms, and equipment repair or return can be costly and challenging in terms of logistics.
Collette Health* (formerly MedSitter) Offers Responsive, Ongoing Support; Respondents Want More Analytics: Organizations utilizing Collette Health* (formerly MedSitter) vary in size, primarily using the solution for virtual patient monitoring. Customers highly appreciate the vendor’s strong relationships and responsive support, particularly since technical issues directly affect their patient observation capabilities. The most frequently cited benefits include a reduction in patient falls and the ability to address staffing shortages and costs associated with one-on-one virtual patient monitoring. However, customers mention limitations, such as the system’s suitability for detecting larger movements (e.g., a patient falling) compared to smaller details (e.g., tubes) due to lower video resolution or small screen size when monitoring multiple patients. Some customers express a desire for more robust reporting and analytics to demonstrate the ROI of virtual monitoring and nursing programs. Additionally, a few respondents point out the lack of integration with Electronic Health Records (EHR) and express a need for improved alerting functionality between observers and responders.
The report also takes a closer look at some other validated vendors. care.ai, established in 2019, was interviewed by KLAS, with feedback from 4 customer organizations, including some large ones. Remarkably, care.ai is the only vendor in the report with all customer respondents utilizing their solution for virtual nursing. While these users are in the early stages of implementation, they report positive experiences so far. Vitalchat, interviewed by KLAS with insights from 2 customer organizations who use the solution exclusively for virtual sitting. Lastly, we have Wachter, a cross-industry vendor with a healthcare division called Wachter Healthcare Solutions. KLAS interviewed 4 out of Wachter’s 10 customer organizations. These respondents are primarily small or midsize organizations employing the solution for a combination of sitting and nursing purposes. Notably, there have been no significant concerns raised about Wachter’s solution to date.
Healthcare leaders don’t have any more excuses for failing to spur change and give consumers not just the healthcare they want, but the healthcare they deserve, Oliver Wyman’s Sam Glick boldly stated during the last day of the Oliver Wyman Health Innovation Summit.
“We know what matters. We know why we are here,” he said. “We know we have an obligation to make healthcare more accessible, more affordable … and something that adds value to people’s lives.”
Speakers from entrenched players, start-ups, private capital, and big tech articulated a vision that borrows ideas from other industries, rides the momentum of breakthrough ideas, and, above all else, puts the consumer front and center.
“It’s not about disrupting. It is about impact,” Tanvi Patel, Director of Amazon Pharmacy, said during a conversation with Yahoo! Finance’s Anjalee Khemlani.
While incumbents may fret over Amazon deepening its reach into healthcare, Patel said the company is no different from any other organization that is trying to transform the industry. Where Amazon is different, she said, is in its ability to remove friction points. Amazon is constantly listening to customers and adapting its services to make life easier for consumers. Home delivery of medication is an obvious example. But she also pointed out the company has worked with drug manufacturers to automate a system for applying coupons to prescription orders, rather than forcing patients, physicians, and pharmacists to scrounge around for them. Amazon wants to expand these types of collaborations and partnerships across the industry to eliminate roadblocks for consumers, Patel added.
Making mental healthcare accessible
Removing friction points is exactly what Gabe Diop is doing to improve access to mental health and substance abuse disorder services. He is the Co-founder of Path Mental Health. Improving access to mental healthcare is personal for Diop, whose uncle died from a heroin overdose.
“We are a healthcare provider group enabled by technology. I picked those words very carefully and that order of words very carefully,” he said.
One of the biggest hurdles that Path addresses is linking patients with a provider who takes insurance. Studies have found that fewer than 50% of psychiatrists take insurance and among those that do, many only accept one or two types. But Path doesn’t just remove that friction point. It also connects patients with specialists who meet their personalized needs, including racial and ethnic considerations. Path typically connects patients with a provider within three days.
Diop said Path is borrowing a page from the Amazons and Ubers of the world and making care more efficient, regardless of the type of insurance a patient has.
Doctors as agents of change
“Don’t think of me as a disrupter, think of me as someone who wants to help your systems evolve and improve care for patients,” Marc Harrison, MD, said.
For much of his career, Harrison was viewed as a disrupter for within the industry, especially during his time leading Intermountain Healthcare where he advanced value-based care designs, extended reach into rural communities, and more. Harrison now wears a different hat as Co-founder and CEO of the Health Assurance Transformation Corp., which is backed by General Catalyst.
Harrison said that private capital, which is often criticized for prioritizing profit over quality, needs to take a different approach in healthcare. Since healthcare changes over long periods of time, the industry needs longer-term investments and permanent capital vehicles.
He also called for physicians to be more involved in spurring innovation and transformation. Jesse Ehrenfeld, MD, President of the American Medical Association, reinforced that sentiment.
“We see innovation being disconnected from our care teams. We see technologies over and over where physician input is an afterthought,” he said, adding that physicians are excited about innovation. One in five physicians uses artificial intelligence, he pointed out. While most of that is currently associated with back-office functions, it helps relieve administration burden and excites them about future uses of AI.
Austin Chiang, MD, is one doctor who is setting change in motion, thousands of followers at a time. Dubbed the TikTok Doc by a number of media outlets, Chiang has nearly 750,000 followers across his social media platforms — TikTok, YouTube, Instgram, and X. On top of his social media presence, he serves as Chief Medical Officer of Medtronic’s gastrointestinal business and Chief Medical Social Media Officer at Jefferson Health.
He uses his status as a social media influencer to counter misinformation and provide consumers, as well as clinicians, with accessible and accurate information about healthcare issues. Hearing from followers who say his posts helped improve their health or provided insights to assist a loved one is tremendous feedback, he said. He urged organizations to create financial incentives to build out physician-led social media platforms and for social media to be added to training programs.
Solving the leadership crisis
Focusing on doing what matters for members is what drives Sachin Jain, MD, CEO of SCAN Group and SCAN Health Plan. Like Patel and Diop, he said healthcare leaders must do a better job of listening to customers and adapting.
SCAN is doing that in a variety of ways. The company in 2022 launched a Medicare Advantage Plan for the LGBTQ+ population. Jain said they expect to have nearly 1,000 members signed up by the end of the year. The insurer is also introduced a new tier in its drug benefit, allowing members to access new therapeutics with an $11 copay.
Even as he highlighted SCAN’s innovations, he cautioned against embracing “toxic positivity” and urged leaders to be more self-critical.
“We do this thing in healthcare where we plant a vegetable garden and we tell the world that we’ve solved hunger,” he argued.
An important question he asks his staff daily is, “If we disappear, will it matter?”
This time last year, Oracle co-founder and CTO Larry Ellison took to the stage at the vendor’s annual user event in Las Vegas and said that his vision for the company’s acquisition of Cerner is to aggregate healthcare data at a national - or even global - level to improve patient care. Ellison’s plans for national healthcare databases would put patients at the center of the healthcare system, he said.
Not only that, but Oracle is prepared to build new data centers to help national governments pursue this. Sicilia said:
In October, we're ready to deploy the national databases for any country that wants to take it. It launches on OCI in sovereign form, meaning the data will never leave the borders of the country that wishes to consume it.
If we have to go and build data centers in the country to do that, we will do that.
It’s a bold strategy, but Oracle appears willing to invest where there is demand. Other conversations with execs have highlighted how it is able to do this cost effectively because Oracle has standardized on its systems, which has allowed for a high degree of automation. This means that building a new data center in any region, according to Oracle, is more cost effective than what might be the case for other cloud hyperscalers.
Sicilia said that whilst Oracle is happy to take a piecemeal approach to the adoption of its healthcare systems, targeting individual buyers, the strategy is at a federal or national level. He added:
I think the federal and national governments are the focus right now. The strategy is that we are going to put a national database in place, but it doesn't necessarily require any of the contributors to that national database to change any of their systems. The strategy is you've got to meet people where they are.
You're not just going to flip out everybody's operational systems and everybody's databases today. It's a federated strategy, where you come in and we federate what they already have in place today.
We did this for the United States government during COVID-19, where we federated all the vaccine delivery on the supply chain systems into a single database and didn't change out anything.
I think that strategy worked incredibly well and then over time, the providers probably will change out their operating systems, because as they feed these national systems, which is an automated process, they'll see some of the new UI, some of the new user experiences, the analytics
Sicilia said that the reason governments may find Oracle’s national healthcare database ambitions appealing is that COVID-19 exposed a lot of weaknesses in the system, where access to real-time data was difficult. The fear of new infectious diseases and the possibility of another pandemic in the future could be a real motivator for a change in how governments think about healthcare delivery, he added.
In addition to this, Oracle’s hope is that because it can take on the entire stack of healthcare delivery, both with Cerner’s electronic healthcare record (EHR) systems, and Oracle’s other backend operational systems, that more demand will flow its way. Sicilia said:
We’re going to solve the whole whole problem with healthcare. The reason that we think others have tried and not been as successful as they could in healthcare, is because they only took on a piece of the problem. They only took on EHR, they only took on billing, they only took on financials, they only took on HCM. We’re going to take the whole thing on.
We have healthcare specific versions for supply chain, for HCM and for ERP, and we're delivering that as a full stack in concert with our electronic health records.
Sovereignty a priority
Oracle is very aware that to achieve its ambitions it needs to take data sovereignty seriously. Data protection regulations are stringent for many governments, but when it comes to healthcare, patients are also very protective over who has access to their data, where it sits and where it can be moved to.
As noted above, Oracle is taking the view that if it needs to it will build data centers to make this workable for its customers. Sicilia notes that data sovereignty is the biggest barrier to adoption. He says:
When it comes to healthcare information, it doesn't matter how big or how small the country is, I haven't met a government yet that’s a fan of saying ‘I'd love to have my healthcare data live in another country’. Even if it's an ally, even it's a friendly
What's been the barrier is that other hyperscalers have been unwilling to invest in building fully sovereign data centers in some countries. What we're saying is we'll build it and we'll be fully sovereign.
Some say ‘we'll have a datacenter for you, but the DR (disaster recovery) is somewhere else’. If the DR is somewhere else, it's not sovereign. We're saying we will build full DR, full data centers in-country, separated by 75 kilometers, or whatever you're comfortable with, with our full stack on top of that. Otherwise, they're gonna have to choose an on premise solution, which is far less attractive.
The other barrier to adoption, as Sicilia sees it, is that because there has been a lack of cloud technology available to healthcare providers, change management requires significant work. Buyers will need to get their head around the fact that the software gets updated automatically four times a year, and training will have to adapt to this. He adds:
As a healthcare provider you need to completely rethink the way you train people and you need to move to a more agile, iterative training process.
Oracle will also be aggressive on pricing, it seems. I put it to Sicilia that another barrier to adoption could be that purchasers of EHR and backend systems may come from different teams and have access to different sources of funding - the strategy isn’t always a unified one when it comes to healthcare provision. He said:
You can still have two different buyers, you can still have different experiences, but wouldn't it be nice to be able to say, we're gonna keep these on common upgrade cycles, we're going to keep these on a common subscription and we know exactly how much we're gonna spend next year?
We can deliver every aspect of what they need and we can give you an all-in price for everything you spend. And generally, it's saying, how many pounds are you spending on IT today? We can do it for X. Even if you're running disparate systems, different buyers, different personas.
Another benefit of this all-in approach, according to Oracle, is that cybersecurity could be improved. Sicilia noted that healthcare organizations are a major target for bad actors - which has been evidenced by a number of high profile attacks - and that by ring-fencing everything into a common cloud perimeter, healthcare data could be safer. He said:
Having a single perimeter defense is far better than having a bunch of different perimeters today. Particularly when you get into the smaller community hospitals, rural hospitals, where you just don't have a lot of sophistication in the IT stack. It's incredibly expensive to go try to ring fence very old stuff, you’re far better off lifting and shifting it to something like OCI.
Asking the wrong questions
One of the examples Oracle showcased at CloudWorld this year was its voice digital assistant, which can be complemented by generative AI. The idea being that a health practitioner could use their voice and say ‘show me this patient’s x-rays and highlight the problems for me’, and then if the patient consented to the appointment being recorded, generative AI could provide a summary of the session for approval by the doctor.
I put it to Sicilia that a lot of healthcare providers are still struggling with much simpler problems - such as the amount of time it takes to turn on a computer and log on - and that the thought of using generative AI or biometric tools is a long way off. He agreed, but said that it’s Oracle’s job to educate on how quickly the technology is progressing and just considering small efficiency savings is a missed opportunity. He explained:
They're asking us to go from 10 clicks to seven clicks, to get that 30% efficiency increase. But as we looked at the problem, and we looked at generative AI and voice navigation, I think we can do better. I think we can eliminate clicks altogether.
I think we go from 10 clicks to zero clicks. They’re asking if we could increase the logout timeout to 30 minutes, to cut down the number of times they have to log on in a day. That saves 10 minutes in the day, which is 10 patients.
We've come up with a system which says: what if we just use your voice as the biometric identifier? What if we basically use a biometric identifier, your voice, and you never have to log into the system again? And then as long as the patient consents to recording, you have persistent recording, and you’re provided with a summary, all the data is summarized using generative AI, and there's a draft for the doctor. Then all the doctor has to say is yes or no.
There’s a disbelief…going from 10 clicks to seven clicks three years ago would have been a very logical request - but I think people need to understand the art of the possible. It’s a remarkable moment. My job has always been in technology, since I left university a long time ago, and I've never seen a breakthrough as big as this with generative AI.
Generative AI is the single biggest breakthrough in terms of productivity for healthcare workers.
My take
Oracle’s healthcare strategy and its ambitions are incredibly bold and very interesting to hear. For those of us that have been following healthcare technology for years, what they are talking about has been a dream of healthcare providers - but is something that has failed to materialize for a number of reasons. None of them really to do with technology. There are numerous stakeholders involved in healthcare provision, which often have conflicting thoughts on how to deliver what’s needed. There are local needs and requirements, which often don’t fit into national frameworks of a ‘desired endpoint’. And, of course, patients are very, very cautious about centralized healthcare data.
That being said, COVID-19 changed the goalposts a lot and there is a renewed desire in the market to do things differently. Equally, Oracle does seem to be wanting to meet healthcare providers where they are and address the different demands head on. The strategy is there, Oracle is ready to go, and now we will wait and see if it pays off. If Oracle is telling us this time next year it has deployed national healthcare databases for various countries, it will be a clear win for the foundations Ellison and his team have laid down. We will be watching closely.
