Taking the “Artificial” Out of AI with Enhanced UX Research

By Alex Ain

The proliferation of digital direct-to-consumer communications in the form of mobile apps, websites, AI, and more has amplified the importance of how consumers interact with and perceive brands.  Along with it, user experience (UX) research has become critical to ensuring that all communication channels live up to the standards of a brand rather than producing conflicting experiences and messages.  

Pictured: an illustrated example of a chatbot, which is a commonly used artificial intelligence tool on websites. The image is trying to demonstrate a poorly optimized chatbot, which does not have the appropriate conversations mapped out and provides a poor user experience.

As market researchers, it is increasingly incumbent upon us to put each individual project in context of the brand and the business and not to only focus on the user experience alone.  Additionally, we need to think beyond the traditional UX measures to serve our client’s larger business objectives rather than focus on UX alone – did the experience accomplish what it was intended to do?  Was the experience intuitive?

A recent GreenBook Industry Trends (GRIT) report suggested that clients and market researchers are at odds as to what it means to be innovative, with consultants often thinking in terms of methodology or platform, and their clients thinking in terms of more strategic qualities such as “knowledge,” “understanding” and “flexibility.” Such is the case with UX research.

The Case of a Clinical Trial Chat Bot

Not too long ago, a HawkPartners pharmaceutical client sought to launch a chat bot (a chat function supported by AI) to match severely ill patients with clinical trials.  The goal was to provide a better process that spurred patients to seek out clinical trials rather than googling them or waiting for their clinician to bring them up.  Instead of waiting for their next doctor’s appointment, they could find the chat bot via any number of websites connected to our client’s brands, answer a series of questions about themselves and their condition, and receive a list of potential trials and contact names to pursue enrollment.

Employing Traditional UX Methods

To answer key questions around intuitive design, overall customer experience, and ability to complete the desired task, we were tempted to rely on traditional UX research alone: setting up a usability lab with desktop and mobile interfaces to evaluate how well patients engaged with the chat bot.  We would track their behaviors and measure how efficiently and effectively the patient could navigate the digital communications channel to achieve a desired action.

Had we stopped at that, we would have met most of the client’s specific research objectives: Was the experience intuitive? Did it successfully match patients with potential trials? The answers to the questions would all have been “yes.”  From a UX perspective, the experience was well designed and extremely efficient in its ability to identify clinical trials.

HOWEVER, patients would never have used the chat bot. That’s because, in relying solely on UX, we (and our client) would have missed out on a couple of critical findings:

  1. Patients don’t necessarily want the “old way” of finding trials replicated. What they want is an easier way, not necessarily a personal or warm experience
  2. In trying to replicate the experience of interacting with a human, the client may have inadvertently reinforced perceptions of pharma as big and impersonal – especially when the bot’s underlying mechanics were exposed through missed cues, as illustrated by this interaction:

           Chat Bot:               What is your name?

           Patient:                  Hi, my name is Julie.  I am interested in finding Crohn’s Disease trials

           Chat Bot:               Hi Julie, what disease have you been diagnosed with?

           Patient:                  I just told you I have Crohn’s

What had been a well-intentioned initiative by our client to connect patients with dedicated and compassionate clinicians would have achieved the unintended consequence of alienating them from the brand.

Using Enhanced UX to Deliver Against Brand Fit

Fortunately, we made the decision to incorporate some more strategic yet basic elements of brand research into the UX research.

By grounding ourselves in the client’s brand positioning and adding a series of questions and projective techniques to explore fit of the chatbot with that positioning, we discovered that the bot led to the perception of our client’s brand as clinical instead of high touch or service oriented.  The client brand took on the characteristics of the chat bot: intelligent and highly capable, yet by definition “artificial” and lacking in credibility.

This insight was gleaned through your basic qualitative research immediately following the UX lab experience.  No additional bells and whistles, just the consciousness to put UX in context of the client’s overall brand strategy to determine (1) whether the experience aligned with the brand, (2) whether it differentiated the brand, and (3) whether it changed attitudes and behaviors.

In the process of lifting the research (and the respondent) out of the UX vacuum and allowing them to reflect on the experience, we discovered that patients would prefer an interface more closely aligned with online shopping.

The Result

Ultimately our client revamped the chat bot to act more like a shopping guide than a friend.  It would help patients select from pre-determined categories such as disease, age, gender, and distance willing to travel, but it wouldn’t attempt to converse or connect.  Subsequent research, again coupling UX with brand research, suggested that the revised bot was not only more efficient in helping patients to achieve their end goal, but in doing so, led to more positive brand perceptions.

It may not have been the highly personal interface originally intended, but the core premise of helping patients, many of whom had exhausted all their treatment options, to find hope garnered very positive brand response, making the investment in a revised approach worthwhile for patients and client alike.

Reach out to Alex Ain to discuss how to enhance your UX research