May 25, 2016

Chatbots Are About to Transform Facebook

facebook profile and thumbs up


We live in a world where instant gratification has become the norm, where marketers must cater around-the-clock to people’s content consumption and brand interaction habits across multiple channels and devices. As messaging apps become the fastest-growing and most-used apps, marketers need to evolve with the way users are choosing to interact, exploring new and relevant technologies — like chatbots.

So what is a chatbot? Chatbots (also: chatterbots) are programs that simulate human conversation, using response workflows or artificial intelligence to interact with people based on verbal and written cues. One familiar example is Apple’s Siri. Another is Microsoft’s Tay, which famously failed on Twitter earlier this year. Last month, Facebook gave businesses the ability to create their own chatbots to communicate with audiences through Facebook Messenger.

Chatbots hold a number of potential valuable uses for marketers, including lead generation, point of conversion, retention and customer service. This POV will address what chatbots are, how Facebook is implementing them, and the opportunities and considerations for pharma marketers.


The concept of chatbots is not new. Back in the days of AOL Instant Messenger, bots like SmarterChild allowed users to have conversations and develop friendships with a machine. At a basic level, they could return sports scores, check the weather and provide movie times.

Since then, chatbots have evolved. Today, you can find chatbots on websites to help with ordering merchandise and answering frequently asked questions. Apple’s Siri is well-versed at responding to queries to provide relevant answers. Facebook’s commitment to chatbots has marketers very excited.

Facebook’s Messenger platform has 900 million users and counting. Facebook-owned messaging apps (Messenger and WhatsApp) send 60 billion messages a day — three times the amount of SMS messages sent globally per day. Clearly, Facebook has the advantage of a strong user base in the chatbot arena already.

Recent Messenger updates, such as prominent page usernames and Messenger codes, allow users to easily start conversations with brands.

We’ve seen pharmaceutical companies use online private messaging tools to invite patients to chat with a nurse about their treatment, handle customer service requests and provide detailed answers to product-related questions. Some pharma Facebook pages even allow for direct messaging.

While these represent pharma’s attempt to be more responsive and patient-friendly, some of these private messaging endeavors quickly became cost-prohibitive due to the amount of required human resources.

Chatbots, on the other hand, provide a new opportunity to revisit what could be positive customer interactions, while automating the conversation. Responses can be fully planned, scripted, and tested to increase consistency and reduce human error. Discussions can be enriched with preplanned images, videos and hyperlinks. And conversations can be fully documented for reference and compliance purposes.

As chatbots become more prevalent, we believe the technology shows promise for relationship building and ongoing marketing efforts, thanks to the following attributes:

  • Scalability — Chatbots can be as complex or as simple as the experience requires.
  • Customization — They can provide personalized experiences in a timely manner based on relevant information.
  • Efficiency — Chatbots reduce the human resources required by distinguishing between simple frequently asked questions and complex inquiries, providing quick, relevant responses to users.

Chatbots can be the frontline of communication between brands and their users. And users benefit as well, receiving more immediate responses to their questions than they might otherwise.

Facebook launched chatbots with several brands already signed on. Users can buy flowers from 1-800-Flowers, receive bank notifications from Bank of America and order food from Burger King.

Chatbots currently provide preprogrammed, automated responses, but have the potential to evolve and become smarter with each new interaction. Rest assured, however, that no newly learned, unapproved responses will deploy and all responses will remain compliant due to the preapproved workflow mechanism borrowed from other pharma social messaging strategies.

As chatbots continue to evolve, it may be harder to tell where the chatbot experience ends and human community managers begin. No chatbot has effectively passed the Turing test, a test designed to tell whether an artificial intelligence is as good at conversation as a human. However, Facebook’s artificial intelligence assistant M, announced last year, is designed to bridge that gap, providing human responses where chatbots leave off. It’s still in testing now but may roll out widely soon, complementing chatbots.


Chatbots mirror one-to-one engagements people have within their own social circles. This allows for a more relaxed marketing approach because chatbots can offer information that feels very much like word-of-mouth recommendations from friends.

The chatbots will complement call centers and interactive voice response systems by reducing some of the volume when human interaction isn’t necessary. For the time being, chatbots will help humans facilitate patient requests, not eliminate them.

In the future, as chatbots become monetized, there will be opportunities for mass distribution of marketing and promotional materials using chatbots, but in the interim, all interactions should be treated as transactional. In other words, the interactions should be user-requested responses that do not require marketing authorization. For now, chatbots will most likely not be used for lead generation or conversion.

Facebook requires that businesses leveraging this technology adhere closely to its policies. Even if users provide a full marketing authorization—or opt-in—they cannot receive advertising of any kind without written consent from Facebook prior to launch of the bot. Chatbots can still be leveraged as a tool to further learn user behavioral tendencies, which allows for a more robust approach to predictive modeling.

Additionally, chatbots will help marketers in the following areas:

  • Acquisition — With the promise to send users reminders and content updates, chatbots can be used as additional entry points into the marketing program.
  • Conversion — By asking questions, we can personalize the conversation to provide relevant information or content in a timely manner.
  • Retention — By encouraging ongoing engagement via continued interaction with timely and relevant messages, chatbots can aid patients in adhering to treatment. With proper permission, chatbots within Messenger allow patients to receive medication reminders, refill reminders or short messages linking to valuable resources.


While some healthcare marketers may be apprehensive about allowing chatbots to converse on their behalf, this new functionality can allow for novel ways of communicating that provide great user experiences. We’ve listed some possible use cases below.

  • Diagnosis assistance — Users may be able to ask chatbots personalized questions about their health. The chatbot then finds resources based on keywords from the user. HealthTap has already made content from its physician network available via a Messenger chatbot.
  • Treatment reminders (adherence) — After users ask questions and interact with brands, the chatbots can offer to remind users to take their medications or refill their prescriptions when they’re running out. This will help patients adhere to their treatments.
  • Appointment scheduling — Chatbots can link directly to portals to aid in scheduling doctors’ appointments. This will make it easier on the patient by eliminating the need to contact their doctor’s office to make an appointment.
  • Doctor and patient advocate location — By providing just a zip code, chatbots can accurately link to or provide a patient with a list of appropriate healthcare professionals close to their location. In areas such as rare diseases, there is an opportunity for chatbots to help patients “find someone like me” in a local patient advocate.
  • Customized education — Chatbots can also help train and educate users on a variety of topics. Setting up automatic quizzes can teach people how to respond to health scenarios like properly using a device or tracking eating habits. With additional input from patients, marketers can better understand content affinity and behavioral paths.


As with any one-to-one interaction, a large concern for people engaging with a brand within a messaging app surrounds keeping these health-related interactions private. Marketers must remain cognizant of user privacy, ensuring it is respected and terms and conditions are clearly disclosed.

Preapproved responses will be vital to the success of chatbots within pharma and, in fact, represent a regulatory-friendly way to answer customers’ questions.

The process for planning chatbot responses can be similar to the way risks are assessed and workflows are mapped out for social media programs. Indeed, pharma is uniquely positioned to develop chatbots, since many brands have already developed extensive response workflows to engage with users on their Facebook pages.

Adverse events will be monitored just like with any other form of communication, and chatbots can be built to follow FDA requirements for monitoring, collecting and reporting AEs.

For example, when chatbots receive user responses that mention keywords such as “hurts” or “stings,” they can trigger a message to report the adverse event to a phone number or email address. Additionally, community managers can monitor conversations for mentions that require a human response.

As an additional layer of protection and expectation-setting, brands can include a description of the chatbot with a small disclaimer and require an opt-in before the user sends their first message. This disclaimer can include hyperlinks to best practices and terms of use for interacting with the bot.

Chatbots can be customized to meet brand needs and objectives. Cost of development will depend on the complexity of the program. The more complex, the more time required to create the chatbot. As of now, there is no additional fee for the use of bots on Facebook.

Chatbots assimilate well with existing marketing and customer service solutions, and social media will not be the only channel that will experience the benefit. As chatbots become more complex, the functionality offers an opportunity for more robust inbound marketing initiatives integrating emails, SMS and direct messages.

One downside to this communication channel is the lack of ability to detect sarcasm or inflection. The coding required for chatbots to understand sarcasm is sophisticated and complex. There is no pattern of recognition that can help detect this type of user interaction yet.


Chatbots can be the frontline of communication for brands, making it easier for them to respond effectively and efficiently to their users. They are an evolving tool that is primed to change the way marketers engage with consumers.

Pharma brands are still regulated, but chatbots can help marketers accomplish their business objectives in new ways. While chatbots may never be able to replace human interaction, marketers should definitely consider developing these tools as a part of their marketing mix.