Could Your Computer Understand People Better Than You Do? | Deloitte US has been saved
Limited functionality available
Imagine a physician—we'll call her Dr. Smith—preparing for an appointment with a patient who is experiencing several health-related challenges associated with her pack-a-day smoking habit and lack of physical activity. Dr. Smith wants to support her patient in quitting the cigarettes and engaging in some exercise. She's never been able to convince the patient to make these changes, but she now has a new partner that gives her hope: an AI-enabled physician's assistant named PAL.
PAL isn't real (yet), but a figment of my imagination. And I'm no expert in artificial intelligence (or health care, for that matter), so my imaginings are truly coming from a beginner's mindset. (For more informed perspectives on these topics, you might check out the work of my Deloitte colleagues who have been inspiring me with their writing about the future of work, of health care, and of AI and cognitive technologies.) What I do know a thing or two about is people and how they relate to and influence one another. For instance, I know we don't all want the same things from our interactions with others, and when you treat people they way they want to be treated, as opposed to the way you want to be treated, your relationships are likely to be stronger. That's the basis of Business Chemistry, after all.
So what's that got to do with my imaginary friend PAL? Well, all this relating we do is very human work. Indeed, it's one of the tasks many believe is unlikely to be taken over by machines or AI, instead remaining a uniquely human capability. And yet, humans are imperfect. We're plagued by cognitive biases, and our decisions are often irrational. Our attention spans are short, and our ability to hold things in memory is limited. Even those who are highly skilled at relationship-building could probably use a bit of help—I know I certainly screw things up fairly often—and lately, I've been thinking about the best way to provide that help. I've been thinking about the power of human-machine collaborations and wondering: What if an AI-enabled assistant knew enough about Business Chemistry to advise us in relating to one another? Whether you are intrigued, doubtful, or a little offended, I invite you to join me in my thought experiment featuring Dr. Smith and PAL.
Imagine that prior to the appointment with the sedentary patient who smokes, PAL pulls up the patient's chart and prepares a preliminary list of questions Dr. Smith might ask, as well as tests she might run. PAL reminds Dr. Smith that the patient is a woman in her 50s and a Guardian, in Business Chemistry terms, then points out that Guardians are often interested in more detail than other patients, and that she should be prepared to answer a good number of questions. (If the patient had a different Business Chemistry type, the reminder might be different; e.g., Integrators often like to start with a little bit more personable small talk, or Drivers typically prefer a very direct approach.)
PAL also reminds Dr. Smith that their analysis of her past patient interactions shows that she (like many physicians) has a tendency to more quickly dismiss the complaints of women than men, so she should be on guard for that particular bias of hers.
Once the patient arrives, PAL listens in to which questions she asks of Dr. Smith and how she answers each of questions Dr. Smith asks her. PAL then adjusts the list of upcoming questions appropriately, urging Dr. Smith to ask, "Is that clear?" after each explanation she provides the patient, because PAL knows that Guardians may stop listening or engaging if they get stuck on something that doesn't make sense to them along the way. (The suggested prompt might be "how do you feel about that?" for an Integrator, or "do you have any ideas about that?" for a Pioneer.)
Once the initial questions are answered and the physical tests are done, PAL suggests to Dr. Smith that she give the patient an extra five minutes to get dressed before addressing future recommendations. PAL reminds her that Guardians sometimes need a moment alone to process information and that giving the patient a few extra minutes now may lead her to think of more questions she wants to ask while she still has time with her doctor—saving them both from a game of phone tag to deal with follow-up questions later. (The suggestion might be to return a quickly as possible for a Driver so as not to keep them waiting.)
As Dr. Smith is getting ready for the recommendations discussion, PAL runs a quick check against a list of common cognitive biases that can affect patient’s health choices and lets Dr. Smith know that the patient’s age and Business Chemistry type suggest she may be most susceptible to the status quo bias (a preference for keeping things the same rather than changing them) and the confirmation bias (paying more attention to information that confirms one's point of view than information that challenges it). (This list of likely biases may change with type and age; e.g., Drivers might be susceptible to the overconfidence bias, thinking they know better than the physician; Pioneers might be prone to the optimism bias, underestimating risks and overestimating pleasing outcomes; and younger patients may be prone to the illusion of control, thinking they are immune to the negative effects of their unhealthy habits.)
PAL informs Dr. Smith that the status quo bias may make the patient reluctant to embrace a lot of change at once, and since she's already juggling her busy career with the schedules of her teenagers and aging father, hoping for the patient to simultaneously quit smoking and start moving regularly is an unlikely prospect. PAL also assesses the patient’s "readiness to change" according to the Transtheoretical Model of Health Behavior Change and determines she is in the contemplation stage for adopting movement activities (i.e., she is considering it), but still in precontemplation for quitting the cigarettes (i.e., it’s not yet really on her radar). Further, based on the symptoms that bother the patient the most and a database of which interventions are most likely to succeed based on which symptoms are primary, PAL recommends that Dr. Smith focus on the movement goal first, starting very small.
PAL also recommends that based on how this patient has reacted previously, it will be important for Dr. Smith to: a) acknowledge the patient's busy schedule; b) appeal to her sense of responsibility to care for others, pointing out that she won't be able to do so if she herself is not healthy; and c) reassure her that the current recommendations don't require drastic change.
PAL then produces:
While Dr. Smith is meeting with the patient to answer any additional questions and share recommendations, she offers a digital tracker that can feed the patient's movement data directly to PAL and that Dr. Smith can review periodically prior to the next appointment. Because PAL knows that Guardians tend to be private, they have already adjusted the setting so that the patient must actively opt in to share the data. Dr. Smith shows the patient how to toggle this feature on and off, but encourages her to give PAL access so they are all on the same page and can work together to get the patient healthy for her kids and dad, as well as for herself. (For a Driver, this might be set to automatically send results to reduce the number of steps required for using the tracker. For a younger patient, the setting might default to a "share my progress with my friends" mode.)
Once the patient has left, PAL prepares follow-up information to be sent, including test results, but also several well-researched, detailed, citation-filled articles about the benefits of even small doses of exercise. (PAL knows Guardians like to read about relevant research, that they want the details, and that they are likely to be paying attention to the quality of the sources.) PAL also includes just one article about the benefits of smoking cessation, with the goal of gradually moving the patient toward the contemplation stage in regard to that behavior. (For a Pioneer, PAL might suggest sending inspiring videos of people reaching hard-won goals, while for an Integrator, they might recommend personal anecdotes from grateful family members of patients who got healthier.)
PAL also drafts a note from Dr. Smith, aimed at overcoming the confirmation bias, which this patient may be particularly susceptible to. The note says, among other things: "There is a lot of information out there that can result in confusion and mixed messages about the right way to get healthy, so I’d like to offer you these articles that our medical team has determined represent the best research to date on the benefits of movement." (The note could emphasize different points according to which likely biases were identified for the patient.)
Upon arriving home and receiving this information and Dr. Smith's note, the patient reflects on how different this appointment felt from past appointments, and she wonders why that is. For the first time she feels seen, understood, and motivated to try to do what her physician suggests. When she receives a quality-of-care survey from Dr. Smith's office, she completes it and gives five stars on the items "my physician really understands me," "I trust my physician," and "I would recommend this physician to others."
For the next several months, PAL receives data from the patient’s movement tracker and sends regular notes of encouragement from Dr. Smith. Because PAL knows Guardians tend to be more reserved and often a bit serious, they adjust the tone of the notes as appropriate—more coach than cheerleader. They keep to the dry, clever type of humor that Guardians tend to prefer, avoiding anything too silly or overly enthusiastic. (For an Integrator, the tone might be more personal, e.g., "I am so happy for you"; for a Pioneer, it might be more inspirational, e.g., "Just imagine where you can go from here!"; and for a Driver it might be "just the facts, ma’am," e.g., "Your tracker shows you’ve been making progress.")
When the patient returns for a check-in appointment six months later, Dr. Smith notices a shift in her demeanor. She seems almost eager to check in, is quite obviously proud of her progress, and is even mildly enthusiastic about trying to do more. She thanks Dr. Smith for helping her get where she is. PAL just nods knowingly from the corner (if an AI-enabled physician’s assistant can nod).
Dr. Suz is a social-personality psychologist and a leading practitioner of Deloitte’s Business Chemistry, which she uses to guide clients as they explore how their work is shaped by the mix of individuals who make up a team. Previously serving in Deloitte’s Talent organization, since 2014 she’s been coaching leaders and teams in creating cultures that enable each member to thrive and make their best contribution. Along with her Deloitte Greenhouse colleague Kim Christfort, Suzanne co-authored the book Business Chemistry: Practical Magic for Crafting Powerful Work Relationships as well as a Harvard Business Review cover feature on the same topic. She also leads the Deloitte Greenhouse research program focused on Business Chemistry and is the primary author of the Business Chemistry blog. An “unapologetic introvert” and Business Chemistry Guardian-Dreamer, you will never-the-less often find her in front of a room, a camera, or a podcast microphone speaking about Business Chemistry. Suzanne is a University of Wisconsin-Madison graduate with an MBA from New York University’s Stern School of Business and a doctorate in Social-Personality Psychology from the Graduate Center at the City University of New York. She has lectured at Rutgers Business School and several colleges in the CUNY system, and before joining Deloitte in 2009, she gained experience in the health care and consulting fields. A mom of two teenagers, she maintains her native Minnesota roots and currently resides in New Jersey, where she volunteers for several local organizations with a focus on hunger relief.