The buzz surrounding chatbots is at an all-time high, but they have yet to be fully embraced by pharma marketers. Due to their relatively low-cost of implementation, ease of use, and omnichannel nature, bots could revolutionize the way we interact digitally. For many, bots are like websites back in 1998—everyone wants to build one, but no one is sure of how and when to do it.
Chatbots are also becoming increasingly popular in large part due to the rise of voice-activated technology. In 2017, 25 million voice assistant devices were shipped, bringing the total number of voice-first devices—such as Amazon Alexa, Google Home, and Siri—to 33 million in circulation, according to VoiceLabs. But bots are not limited to voice interactions. Gartner Research states that 85% of customer interactions with brands will be managed without a human by 2020. Also, according to a Facebook survey, 56% of people would rather message than call customer service and 53% are more likely to shop with a business they can message.
Although still relatively new in the space, there are a number of healthcare-specific bots that have already hit the market and are making an impact.
Woebot engages users to help decrease anxiety and depression through natural language processing and therapeutic expertise. Although this does not replace the need for psychologists, it helps consumers talk through issues they experience and guide them toward a solution. Seattle Reproductive Medicine used chatbots to simplify filling out the application for an egg donation. And Your.MD is a bot that provides relevant and trusted information to people experiencing health problems. It allows them to get to the root of the problem quickly and solves for smaller issues that may not require going to a doctor’s office.
Building Your First Bot
There are a few key considerations to consider when designing and building a chatbot. To start, the chats should be simple and use structured input when possible. Especially when it comes to healthcare, the tone should be straightforward, professional, and informative. And although the chats should be empathetic in nature, the bot should not pretend to be a human. Instead, it should offer tips for users to get started in addressing their query and provide feedback from there. Also, the answers should be more than text, they should include images, videos, or even trigger the execution of a task.
It’s important to understand that chatbots are not one-size-fits-all. You should determine the end goal of this experience and build a bot that will best help you get there. The four main types of bots to consider are:
- A Directed Bot uses an interface with predetermined menu options. This benefits the consumer by making it easier to know what to ask.
- An Open Bot is similar to having an online conversation and allows free text entry.
- A Hybrid Bot combines free open text entry with suggested options (or buttons). This is the most popular form of bots.
- A “Live” Bot blends AI conversations with human interactions for specific questions. If a question includes a specific topic, a trigger will direct the conversation to a live representative.
Of course, the primary barrier to bot adoption in pharma is the infamous two words: “Adverse Events.” How can the bot handle these?
By teaching the bots the right keywords and building the appropriate logic, bots can effectively handle AEs and get approved by your MLR team. It is important to reinforce that bots are not like Siri; they do not scout the open web to provide answers. Only the pre-approved answers will be used and nothing else.
Bot creation is very different than web or app building. There is no UX, no UI, and no conventional creative. Instead, the focus is on the scope of the bot, the intents, the answers, the tone and voice of the bot. Also, bot development needs to be 100% agile. Everything does not stop at launch. Using bot analytics, we can identity the top requests, the top keywords, and most importantly, the emotion of the user. Using this novel type of analytics, we can add new answers, refine the scope and always optimize to make the bot as efficient as possible.
We are just scratching the surface of what we can do with chatbots in healthcare. With AI and voice technologies becoming more pervasive, cheaper and easier to adopt, we can expect to see many new use cases and applications of the technology, and, ultimately, real-world results.