Companies in a variety of industries are leveraging conversational artificial intelligence to better serve customers and support employees. Some healthcare provider organizations are using conversational AI to help caregivers and patients.
The technology is being used in healthcare for such things as patient chatbots, medical triaging and patient support, patient data management and access, and access to medical services.
Healthcare IT News recently sat down with Manu Aggarwal, a partner in Everest Group’s business process services practice. Everest Group recently evaluated 20 vendors in the conversational AI field and classified them as leaders, major contenders and aspirants in the firm’s PEAK Matrix framework.
Here, Aggarwal explains the workings of conversational AI and reveals some of the findings of his firm’s vendor evaluation.
Q. Why should conversational AI be important to healthcare CIOs and other health IT leaders today?
A. Among all industries in the world, healthcare is the one that’s all about people – people such as doctors and nurses who are helping other people, patients, get better.
Despite people playing such an important role in the industry, healthcare is not known for being savvy in terms of patient experience. Industries like e-commerce, retail, even banking have gone through their own waves of material innovation, but healthcare has yet to replicate the similar experience-first culture.
One of the main reasons why the people-centricity hasn’t caught on is the inability of healthcare technologies to speak with each other. We continue to operate in an environment where your medical records could be lying across five different EHR systems, all your enrollment and benefits data could be with your payer who doesn’t necessarily communicate seamlessly with your pharmacy, and in case you have to go out of network, or worse yet, out of your home state, you are in for a whole new experience.
In such a disparate environment, conversational AI helps both payers and providers manage their patient data and experience better. Healthcare as an industry has been making steady progress in implementing digital assets such as automation and analytics for years now, but most of these are considered a means to an end.
For example, a way in which an insurance company can perhaps connect two data points or help process a claim faster. Conversational AI helps bring these solutions together and unlock their full potential where it truly matters the most – patient/member interactions.
Q. How are healthcare provider organizations leveraging conversational AI to better serve patients and staff?
A. There are many ways in which providers today can leverage conversational AI, but there are three most common types of use-cases.
Medical triaging and patient support. Imagine a patient who is experiencing some light symptoms and wants to access some medical support to identify whether their condition requires an in-person visit. For years, providers have been offering triaging support via telehealth channels. In fact, during the COVID-19 pandemic, telehealth likely became the default way for many patients to get medical attention in non-serious conditions.
Conversational AI can help pick up key symptoms during a live patient conversation and help the nurse who is triaging the case identify potential associated medical conditions faster. Moreover, with access to these tools, the system can also dig into a patient’s medical records in a matter of seconds and identify any patterns that may point toward a more serious condition.
Another way conversational AI can help in such situations is detecting speech patterns. A patient who is really nervous or anxious can give off subtle signals, which can be detected by such a tool, which eventually signals the same to the nurses who can help the patients calm down.
Similarly, if a patient is not familiar with the medical jargon, the tool can detect the confusion in speech patterns and suggest alternate terms to the nurse who is managing the conversation to help explain the associated medical condition better.
Patient data management and access. As mentioned, patient data often resides in various different silos. Another way conversational AI can be helpful is by helping patients access that information faster.
This could be a simple use-case of refilling a lost prescription or something more complex such as releasing test results or immunization records to a new facility on a patient’s request. Further, conversational AI also can help improve security measures and prevent identity theft through technologies such as voice recognition.
Access to medical services. One of the most important reasons a patient might call a provider is to book an appointment. Be it for an annual medical check-up or to consult a specialist, everyone goes through the same scheduling process.
However, patients must often rely on their own research to identify which facilities are best known for which type of care. Further, patients are often unaware of their coverage status until they arrive at the provider’s office.
Conversational AI can enable several such use-cases for patients, starting from eligibility checks, recommending best doctors and facilities for specific specialties in a patient’s neighborhood, and even helping with scheduling time at the appropriate facility.
Additionally, the tool also can enable more seamless operations at the provider’s end, enabling providers to get timely intimation of a patient’s visit, get them seamless access to patient’s records at the right time, and improve the overall process of admitting a new patient, all the while ensuring that their core infrastructure is appropriately utilized without the need to create long backlogs.
Q. Everest Group recently evaluated 20 major vendors in the conversational AI arena. Which vendors have the vision and the capabilities to deliver innovative solutions and make a significant market impact?
A. Conversational AI is an evolving space where we are seeing vendors making investments to innovate and improve their products. The leading investment themes where significant investments have been made include:
- Improving the conversational capability of the solution by using more sophisticated natural language processing and AI models to help improve accuracy and context understanding, allowing context switching and handling multiple intents.
- Improving agent-assist capabilities to ensure better compliance, offering next best action, and quicker post-call work.
- Improving analytics for better customer profiling and segmentation and also to offer real-time insights.
We see a number of players making good progress on all these fronts, such as Amelia, 24.ai, Kore, Onereach and Avaamo. The investments have also resulted in good success for these vendors, with all of them posting healthy growth in 2020.
Avaamo, Kore, OneReach and Amelia also have a good experience of working with the healthcare industry.
Outside the vendors analyzed as part of this research, Nuance (now acquired by Microsoft) also has advanced conversational AI capabilities and a strong play in the healthcare market.
Q. What is the future of conversational AI in healthcare? What should healthcare CIOs and other health IT leaders keep their eyes on?
A. In all honesty, the healthcare industry is a bit late to this party. Industries like BFSI (banking, financial services and insurance) and telecom already have been reaping the benefits of conversational AI for several years now. In fact, in the 2019-20 period, spend on conversational AI grew by almost 80%, primarily driven by those industries.
However, when it comes to technology, there’s always some benefit in being a second-mover, too. The tools and solutions in this space already have gone through years of rigorous testing and are now readily getting deployed, often in almost an out-of-the-box manner. Further, a lot of new features and functionalities have already been incorporated, too.
Having said all that, as with any AI, the core value gets unlocked only when the engine is allowed to learn. Since healthcare is a late adopter, a lot of healthcare-related data has not been fed to such models as compared to what other industries have done. Data sources, ranging from previous chat logs to incoming emails from members and patients to any other communication channel, all form a great data asset through which this AI engine can learn exponentially faster.
Another way these AI engines work is by observing patterns that humans follow. And more so in healthcare than in any other industry, context is really important. How to treat symptom descriptions, medical terminologies, clinical protocols, etc., are all things that the engine will need to be taught so it can get better and more efficient in the way it processes.
Finally, with a huge rise in devices and personal health assistants, there’s a corresponding rise in valuable data getting generated from those sources too. Feeding such data to the conversational AI engine can extend the benefits of this tool to not just be reactive anymore but also enter the realm of proactive care.
Detecting fluctuations in any vital statistics can alert both the patients and, if appropriate, their primary care providers much ahead of time in terms of any meaningful need for intervention or preventive care. Conversational AI can help enhance the health outcomes for all patients in manifold ways and truly unlock the healthcare vision of “caring about and for people” as opposed to “treating the diseases.”