Saint Elizabeth Health Care’s chatbot offers enhanced family support

Family caregivers, meet Elizzbot. She’s your sounding board, your resource, someone to vent to, and a source of constructive solutions when you’re burned out.

She is, in her way, a daughter and a granddaughter, tracing her family tree to the first chatbots. But the project of Elizz, the online brand of home-care provider Saint Elizabeth Health Care, is very much her own bot.

Elizzbot is still to be refined before her imminent rollout to Saint Elizabeth staff before being made available to the general caregiving public.

“We know Elizzbot is unique because she was designed for family caregivers,” says Allyson Kinsley, senior vice-president of strategy for Elizz, Saint Elizabeth’s online brand, launched in 2016. The brand’s persona permeates the bot.

“We had a personality we wanted to come out for the brand,” Kinsley says – optimistic, curious, smart, vital, gutsy, persevering, quirky.

Chatbots – computer programs that interact with users in a near-human way, thanks to various artificial intelligence technologies like natural language processing, pattern recognition and neural networking – are quickly displacing apps as the tool of choice for interacting with online brands.

Research firm Gartner predicts that chatbots will power 85 percent of customer interactions by 2020. By then, the average person will have more daily conversations with bots than with their spouses.

Already, in fact, 27 percent of Americans aren’t certain whether their last customer service interaction was with a chatbot or a human, according to PricewaterhouseCoopers.

While 80 percent of companies have bot plans on their radar, healthcare is a few years behind, says Zayna Khayat, a future strategist who joined Saint Elizabeth in February.

But it’s a field ripe for bot-based innovation. Bots can serve as medical assistants, both personal (managing appointments and medication compliance) and clinical (smart triage). There are condition-based bots for cancer patients, palliative care, and diseases like Chronic Obstructive Pulmonary Disease (COPD) – the breatheAgain bot from Toronto’s University Health Network being one example.

Bots can be aimed at particular demographics (Eve for women’s issues, Vivibot and emojiHealth for teens).

In Japan, cross-bred chatbot/physical robots – Paro, Jibo, Pepper, Autom – serve as companion robots in most senior citizen homes, says Khayat. Bots serve up reminders, answer questions, and guide users to online resources relevant to their condition or situation.

Elizzbot can trace her lineage to the first generation of chatbots. In 1966, Massachusetts Institute of Technology professor Joseph Weizenbaum built ELIZA, a computer program that could engage in dialogue with a human, responding to “patients” with scripted, non-directional questions. (For example, ELIZA would ask, “How does that make you feel?”)

The system was built by parsing input for weighted keywords and using pattern recognition and substitution rules. Though ELIZA was a parody of a Rogerian psychotherapist, test users quickly began to credit her with humane attributes like insight and empathy.

But Elizzbot is more directly related to Tess, a psychotherapy and psychoeducational chatbot engine developed by Silicon Valley startup 2XAI, which specializes in using AI and automation to expand mental healthcare services.

The startup was founded in 2014 by Eugene Bann, a programmer working on sentiment analysis algorithms when he met partner Michiel Rauws. The pair developed a counseling program, Karim, to help Arabic-speaking Syrian refugees at a camp in Jordan, where mental health issues were rife and resources to help cope virtually non-existent.