Intouch Solutions’ Sean Hartigan and Heartbeat’s David Sakadelis say their pharma clients are skittish about the use of chatbots and AI, with an understanding of these technologies more influenced by the grim future of Skynet in the “Terminator” movie franchise rather than the reality of where AI is right now.
“The pharma companies, they don’t fully understand the technology yet, they’re not quite there yet,” says Sakadelis VP, group director, head of technology at Heartbeat. “But I do see some movement there, they finally understand that native apps don’t see a lot of adoption or a lot of utilization. So they think about chatbots and where they fit in. Unfortunately I don’t think they really understand technologies like AI fully. There’s some apprehension to it, they’re not sure how this works from a regulatory perspective, and the other thing is, when they are touching on these new approaches, to voice and AI, they’re pigeonholing it into, ‘This is an augmentation to live person support,’ rather than something that can connect better to the patient and drive things like adherence.”
According to Sakadelis, Heartbeat does a lot of knowledge sharing with clients about new technologies. “We have a program here, Experiential Technologies, we call it Ex for short,” he says. “It’s a lab program. The business problems don’t change in pharma. There are a lot of typical problems that we are always trying to solve through our campaigns and our solutions. Technology is changing and constantly evolving, so we try and connect new technologies or something that has evolved to a business case or problem for our clients, and we do some experimentation around that to see if it solves a problem.”
Through Heartbeat’s connections to non-health Publicis agencies such as Sapient and Digitas LBI, Sakadelis gets to talk with non-health creatives and marketers about what’s happening in the consumer space and how that can be brought into pharma.
“Our opportunity to introduce that with our clients is through our lab program,” Sakadelis says. “We can just throw something against the wall and see if it sticks. We can talk about virtual reality and augmented reality. We bring clients in and they love to engage with this technology, and we try and help them connect the dots in ways that lead to solving their problems in new ways, through new technologies. More often than not, it doesn’t lead to some campaign. It’ll be a very watered-down version of it. But we keep pushing it. We’ll keep moving the needle on it.”
One of the programs Heartbeat did was centered around chatbots, where the participants tried to tackle challenges such as adherence and patient support through AI, machine learning, and chatbots.
“What we found is, on the surface, everyone heard of AI and chatbots, and think that the thing [AI] is off and learning on its own,” Sakadelis told Med Ad News.
Part of that apprehension was the publicity about how Microsoft’s Tay the Twitter chatbot was trained to say racist things within 24 hours. Of course this made pharma executives fearful about using chatbots.
To counter that fear, Heartbeat’s workshop participants did an experiment, centered around creating a simple chatbot for registration purposes, “but we wanted to have more than sort of a multiple choice approach, which pharma is taking on quite a bit because that seems safe,” Sakadelis says.
In a multiple choice approach, the chatbot says something, and the human can only respond in one of three ways. “We wanted to do more than that, we wanted to incorporate natural language understanding, this ability to type in something and the system interprets it into what is called an intent,” Sakadelis says. “Through this experiment, we learned that there is a couple of ways of doing this, and there is a way to keep it a bit more open-ended where the system learns from things that folks are saying to it, but there are also ways to cap that and control it a bit more.”
Using this method, chatbots can be trained on very specific terminology that can be translated to very specific intent.
While some limits on AI and chatbot learning are needed, “There’s a fear that if it becomes so guided and so scripted, the user experience really suffers,” Sakadelis says, “And if you’re looking to use these technologies to create a more humanistic approach to patient engagement, you kind of fail in that way.”
AI has come a long way, but is nowhere near “Terminator” awareness. “A lot of this is still just really pattern recognition,” Sakadelis says. “It relies on the quality of the data that we’re using to supply the logic to the machine learning so that it can make recommendations. There are a lot of companies at the forefront of this, like Google, which just has the mass of data needed to support more intelligent systems, but as a whole, AI is not really there yet.”
Despite the fears and the tech barriers, patients – pharma’s customers – want chatbots.
“People want that frictionless experience, they’re getting it from retailers, they’re getting it from everyone outside of pharma,” says Hartigan, senior VP of strategic planning at Intouch Solutions . “So on our mobile devices, we kind of want the same thing from pharma. We want that simple, intuitive, all-about-me experience that a chatbot or a voice assistant can help us with, in lieu of maybe even having to talk with a human being because that can get uncomfortable, because we love our anonymity.”
To try and get pharma clients to the point of trying out what consumer companies are already doing in mobile marketing, Hartigan says cross-functional teams must be involved. “That’s marketing and sales and med affairs, and technology, and even the C-suite,” he says. “If the C-suite doesn’t have this vision, for a completely customer-centered journey that we’re going to provide to our doctors and patients through digital and through the use of data, on phones, everywhere, if the C-suite has not bought in, forget it.”
Hartigan says any tech initiative in pharma must be introduced in steps. “Pharma loves to do things in phases, and it’s probably the best way to go is do pilots, and introduce a new way of doing things through this cross-functional team effort,” he says. “And then once people see it’s not scary, then you can expand what you’re doing and do more of it. But you have to start small.”
The medical, legal, and regulatory executives also need to be brought on board early.
“The ultimate thing is to get that cross-functional team and the C-suite to a line, and then once that happens, you can do your creative review with your MLR and say, ‘Look, all these teams looked at it. We looked at it from risk mitigation, we looked at it from complexity, we looked at it through all these lenses that we have to consider, and we have a plan that we know that will work and will mitigate risk,’” Hartigan told Med Ad News.
When informing its clients about new technologies, Intouch, like Heartbeat, regularly gets all the potential end users together – including regulatory – for innovation labs.
“You shut the door and you put the chains on it, and you make people fight it out for what they can do vs. what they can’t do,” Hartigan says. “And it does work, we’ve done it for large pharmas for everything from clinical trial programs to just branded programs for both patients and doctors. It’s collaboration, communication, goal-setting that everybody’s aligned to, and then you can do the risk mitigation with regulatory.”
The regulatory executives must be included from the beginning of any project. “Don’t just bring it to them, and say, boom, here it is, this is what we want to do,” Hartigan says. “Then they’re like, ‘Whoa, wait a minute, you crazy dreamers!’”
Pharma must also bring in the constituents – doctors, patients, pharmacists and payers – “because at the end of the day, we’re talking to ourselves. If we just do it with our internal client people and the agency people we have to do it with some of the people we’re doing it for. And then if regulatory can see the end users chiming in to see the value in this forward-thinking stuff we’re doing, they’re more likely, of the risk-mitigation makes sense, to approve a pilot,” Hartigan told Med Ad News.
Because privacy concerns are already being expressed by many consumers, pharma has to be even more stringent. “We’re trying to make sure that we’re really transparent and that our privacy policies about data collection are something that consumer and public can get comfortable with,” Hartigan says.
There are already advancements in biometrics that can help assist with keeping mobile data private, such as voice recognition and facial recognition, signature recognition, and fingerprint recognition. But biometrics can also be used in interesting ways in mobile campaigns themselves.
“For example, dermatological products in pharma might make use of facial recognition for something,” Hartigan says. “And signature recognition right now is being used for sampling by reps and doctors. All of these new biometric technological features can be added more and more to pharma as we go.”
And to Hartigan, “the sexy part about mobile going forward is being more surgical with what we’re learning and prescriptive back in speed to market, back to these people as they’re engaging with us on mobile.”
Heartbeat has built a proof-of-concept mobile messaging service for people with diabetes, Sakadelis says. The service would be something that is prescribed by the physician at the point of care through EHR system, using verbal consent.
The service pulls some information about the enrolled patient out of the EHR – the IC10 code, the patient’s first name, and their phone number – and by engaging with the patient through SMS, the service directs the patient to a HIPAA-compliant interface where they can engage with a virtual assistant.
The virtual assistant then can make recommendations about diet, exercise, nutrition counseling, and where patients can find these support services. The suggestions are made based on the patient’s progress on how they are doing with their condition.
“And we can offer those value-added services to each of the pharma brands,” Sakadelis says. “If they want a program to support the patient through patient advocates, we can offer to connect the patient at the right moment. Or human in the loop support for patient advice, or getting them to their next appointment with Uber Health.”
Presently, Heartbeat is looking for a health system to partner with on the diabetes messaging chatbot. “Ultimately we see this supporting the health systems and pharma,” Sakadelis told Med Ad News.