How we use artificial intelligence at Wellzesta

Posted by John M. Robinson MD, PhD on March 12, 2019
final man on train-2

Artificial intelligence can benefit seniors’ health & wellness - here’s how.

Data Science, a relatively new interdisciplinary branch of study, has produced a set of related technologies: Machine Learning, Artificial Intelligence, and Predictive Analytics. You may have heard of these, but you may not be familiar with what they are, or, more importantly, what they can do. There are many evolving applications of artificial intelligence (AI) in seniors’ lives. This article introduces these technologies through the features that they enable in Wellzesta’s senior living software.

How it works

An important part of Wellzesta Life is a wellness coach named DOT. DOT, which stands for “Do One Thing,” is powered by AI and has a variety of skills that give “her” human-like qualities.


DOT can speak information about a community’s posted events, dining specials, and more. DOT’s human-like voice is powered by a neural network provided by Google. Text is fed into the neural network (input), and the neural network produces an audio recording (output).

The neural network is one of the better known machine learning algorithms. Neural networks are fairly simple constructs. They begin without any “wiring,” then, like the human brain, rewire upon training. As we designed DOT, we examined the performance of several neural networks that convert text to speech. To us, Google’s network produced speech with the most life-like qualities in terms of tone, inflections, and prosody.

Wellness Coaching

DOT possesses a cadre of bots that perform an ever-growing number of skills.

What is a bot? Let’s contrast robots, which are familiar to you, with bots. Robots are physical entities that usually integrate several skills such as vision, intelligence, and movement.

Bots are software entities that perform just one skill. They are autonomous agents that have one (or more) inputs, one goal, and one specific skill (output). For instance, Wellzesta utilizes a birthday bot with one skill: to issue a birthday card to a person on their birthday. Bots can be long-lived or short-lived.

Bots can be scheduled to do a task (“scheduled-bot”) or called up on-demand (“called-bot”). The birthday bot is a scheduled-bot. Every day, at midnight, the birthday bot is launched. In less than a second, it sifts through the birthdays of all users, sends electronic birthday cards, then goes to sleep. Coach bot is another scheduled-bot. The coach bot analyzes the user’s accrued points in the eight dimensions of wellness, then makes personalized recommendations.

The devotion bot is a called-bot. When the user clicks on a Bible verse in the daily devotional article, the verse bot is called to translate the chapter and verse into a format that the bible app understands. The implementation of the spiritual wellness devotion bot is described here.

Each bot uses AI to turn input into output. Basic examples of AI are translation (also called mapping), simple arithmetic, comparisons (also called thresholding), and look-ups. More complicated logic involves data reduction and modeling, including principal component analysis, non-linear parametric estimation, and reduced models like hidden Markov models.


Predictive analytics uses historical data to predict some consequence. In Wellzesta Health, we use predictive analytics to assess health risks. For example, we could predict the following: “If you have no friends then you are at risk for social isolation.” Predictive analytics leverages statistical correlations to make Bayesian-type statements such as “You are more likely to ….” and “Your risk of developing diabetes is ….”. You guessed it: DOT relies on bots that perform predictive analytics to assess health risks for seniors.


Wellzesta DOT uses technologies from Data Science - Machine Learning, AI, and Predictive Analytics - to deliver spoken content and perform personalized health and wellness coaching for seniors. With these tools, Wellzesta is able to offer a higher level of individualized, accessible care for members at our partner communities and look toward the future of artificial intelligence for the senior wellness industry.