Incorporating AI into Telehealth

Tonya Hall: I’m Tonya Hall and joining me is Evan Kohn, Chief Business Officer at Pypestream. Welcome, Evan.

Evan Kohn: Hi Tonya, thanks for having me.

TH: Of course. Tell us, what does Pypestream do?

EK: Well, Pypestreams’ an all-in-one cloud messaging platform that large corporations and health organizations use to engage with consumers, as well as patients, at scale, with conversational AI. So, our company was born out of the notion that people hate calling call centers or sending emails to customer service only to hear back in days, so the combination of automation and messaging is allowing organizations to become always-on entities. So, this includes hospital systems, healthcare providers, insurers, because the days of 9-5 customer service hours just can’t be over soon enough for the average person who wants to engage with a business or organization, really, on their own time; in the same way they engage with friends and family. So, you know, people today expect immediacy and messaging is really the fastest path for organizations to operationalize this – I’d add that these types of investments are paying pretty enormous dividends for companies and organizations today. Many of our Fortune 500 customers are seeing 30% plus increases in customer satisfaction, but also 90% plus cost savings compared to anywhere from $6-$15 averages per call for many companies to field basic questions from their consumers, from their audience. So, customer-centric enterprises today can really embrace this move towards messaging because it doesn’t compromise customer or patient experience, it’s ultimately transforming it as part of this ongoing paradigm shift.

TH: With the shutdowns and remote relocation related to COVID-19 how is the use of chatbots and conversational AI in telemedicine evolved?

EK: Yeah, we’ve seen a rapid increase in interest from health organizations, wanting to accelerate their digital transformation efforts. To date telemedicine has been assumed to only be conducted via phone or video and we’re now seeing this consumer shift toward messaging enter into the health space, as well. You know, for example, early in the COVID-19 outbreak we partnered with an organization called MedCall Advisors that has hundreds of ER physicians around the country, and we worked with the to automate coronavirus screenings, specifically conversationalizing a series of questions to determine if a patient is at risk of having coronavirus. So, those questions according to CDC guidelines, we found that to be a much more accessible way for patients to find out if they should be talking to a doctor or a nurse, as opposed to navigating a clunky web portal or filling out an onerous form. So, this automation of COVID-19 screening via messaging allowed us and MedCall to determine what population of patients should be escalated to a care coordinator and ultimately a doctor. And this ended up saving a lot of time for doctors and nurses, so it allows us to honor their precious time, especially amid the outbreak as there’s so much dependency day to day on those frontline workers. So, these types of messaging experiences are also helping ensure accurate, consistent, quick dissemination of information from health organizations to patients. And as health organizations are spending billions on even basic customer service questions in the US, these new opportunities as telemedicine capabilities evolve beyond phone or other means into the messaging paradigm has really become a no-brainer for a lot of health organizations. And also because, you know, the use cases are really unlimited in terms of what you can automate through conversational AI. Everything from medication monitoring to proactive appointment alerts, sharing lab results, just keeping patients up to date on their condition or helping them with care navigation. So, that’s something we spend a lot of time on as well, our team’s working with health executives to really prioritize what are the use cases that would pay the highest dividends in terms of operational efficiency for health organizations to operationalize, but also what will pay the highest dividends in terms of patient experience and really improving outcomes.

TH: So, to what degree is conversational AI spontaneous versus pre-scripted?

EK: Sure, so it’s really a combination. So, once an organization sets out the use cases to automate we have a set of pre-designed conversation flows, pre-built backend integrations as well as training data specifically for AI in healthcare. So, we’ll take that and implement it according to those set use cases, but it’s not purely a journey where you’re navigating carousels or watching videos or tapping on option buttons within a web conversational interface, there’s a free text entry field as well and that’s where the power of our AI, it’s called Dovetail, will parse the user’s intent. We take a bit of a different approach to that, as opposed to doing basic keyword match or using a single classifier like many organizations do, we’ll use a variety of classifiers and use different combinations of them according to the use cases, the industry, this specifically – in this case of healthcare – and on top of that use semantic search, as well as sentiment and tone analysis and in that is even emoji analysis; so, really adapting to the most modern and informal ways that customers and patients might interact. So, we’ve found rather than where a lot of providers will just become obsessed with the natural language understanding component of these types of solutions where it’s all free text entry whereas, on the other end of the spectrum, there are other providers that just do basic chatbots that are purely rule driven. We’ve brought those together and through the design exercise of really tailoring an experience to specific use cases and ultimately the business goals of this health organization we’ll bring those capabilities together for a unique patient experience that is ultimately effortless and is going to drive a higher satisfaction and really offload the mundane components for agents, for care coordinators, for doctors and nurses, around data collection or the basics that automation can really help with.

TH: Pandemic or not, telemedicine has a bright future, right? So, where are the next big opportunities for chatbots and conversational AI in the space?

EK: Yeah, well let me first say, there’s a really wide spectrum now between a basic chatbot that you can build in a university dorm room in two hours and the other end of the spectrum, you know, advanced, secure, scalable, HIPAA compliant, PCI compliant conversational AI. And that’s where, we at Pypestream, have really dedicated a ton of product and engineering resources toward that latter, so that ultimately, we can automate more experiences and offload the manual work from organizations in a preferred way around messaging for consumers. But, there are new use cases that we’re seeing everyday brought to life, and I think this paradigm shift is really making health organizations more resilient by allowing their workforces, including doctors and nurses on the frontlines, to dedicate maximum time toward patient care while that mundane data collection is handled by AI and automation. So, I think that we’ll increasingly see, you know, with the advent of 5G, and even more so, increasing consumer expectations toward immediacy in their lives. These types of telehealth are going to grow and first, maybe for rural organizations, but also whether or not the pandemic is here for a few more months or another year or two, this has already had a major impact in terms of just shifting the mindset, I think, of a lot of healthcare executives and they’re now in the mode of thinking what are the use cases that will really help us serve patients better through these new means of communication.

TH: Evan Kohn, Chief Business Officer at Pypestream. If somebody wants to connect with you, Evan, how can they do that?

EK: Sure, well they can go to pypestream.com or they can DM me on Twitter @evankohn

TH: Thanks again, Evan, for your time. And find more of my interviews right here or at tonyahall.net. Thanks for watching.

This interview was originally published on ZDNet.