Behind this AI is the woman shaping its future: Meet Sara Hines, Humana
Would AI exist without women?
Of course not. Effective AI is built by a diverse team—one that is inclusive to women as well as individuals of any race, gender identity, sexual orientation, belief system, and even accommodating to variations in how one thinks.
It has to do with empathy. AI is an artificial life brought into the world; it is our child—in fact, at this current moment in technological advancement, AI is a child. In the age of infant AI, data architects perform what’s akin to watching their AI toddle around its environment and bang into walls. They desperately try to cushion the sharp edges and impart AI with essential knowledge to improve.
At the same time, an AI tracks every move and mention of its creators. It absorbs the framework of their thoughts. But this also means copying the flaws wholesale. Unless we ensure AI gains the perspectives of many different people we risk embedding harmful biases and continuing a cycle of systematic inequality.
In today’s companies, women leaders are playing a pivotal role in bringing inclusive AI into the world. IBM has recently launched its inaugural IBM Women Leaders in AI in recognition of women advancing their companies journey to AI across diverse industries around the globe. In this blog, I’ll spotlight one of these leaders who is shaping the future of AI.
How to build AI you can trust – Sara Hines
“We will never try to introduce AI as though you're talking to a human, because I think that continues to build trust… It's like the car, the telephone, electricity; at first, people are cynical of it and they're fearful of it.” - Sara Hines, Humana
Sara Hines is the director of provider experience and connectivity for Humana, the third-largest health insurance provider in the United States (based on GAAP consolidated revenues among major public managed care companies in FY2018). Sara works closely with healthcare providers such as your doctor or other medical professionals who take care of you.
From her early beginnings with Humana, you could say Sara was on a mission. She was recruited onto a small “swat innovation” team, with few guardrails other than to bring “disruptive and transformational technology to the provider service experience realm.” With half a career spent in IT and the other half in business strategy, Sara built up strong acumen in both and quickly began to deliver on that directive.
Sara identified that Humana fielded over 1 million provider calls every month. These calls were very different than member calls inquiring about insurance benefits and account information. Provider calls were transactional; the callers were educated and had their intents firmly in mind. Sara quickly saw an opportunity for using an AI-powered virtual agent to revolutionize the contact experience for providers. But she did not begin down this path without some trepidation.
She recalls one encounter of talking with a family friend at a birthday party. They struck up conversation about Sara’s involvement with AI. The friend confided, “Sara, I am really scared about [AI].” Sara was surprised by the reaction: “He was genuinely concerned. It took me aback because I thought he was almost teasing me. But he wasn't.”
Experiences like these caused Sara to view AI cautiously, but still she knew with the right approach it could be wisely used. “There is a set of responsibilities that come with AI… there needs to be a very focused awareness of how people perceive it and I think there's a responsibility that goes along with that,” said Sara.
Her core principle was simple: always be forthcoming that the caller was talking with an AI, not a human—according to Sara, this crucial transparency builds trust.
Humana virtual agents find their voice with IBM Watson Assistant
With AI aspirations, Sara and the Humana team began putting the plans to develop their use case into motion. But the course of actualizing AI into the real world never did run smooth; before long, a major hurdle emerged.
Since providers frequently called Humana over legacy desktop phones, the virtual agent would need to leverage narrowband technology, not broadband as typical for services like Alexa or Siri. This proved a major limitation and bottleneck to getting the AI project off the ground. Sara quickly learned that her use case was virtually “nonexistent.” In other words, no one had attempted it before.
So began Humana’s journey to AI. It wasn’t easy: full of false starts, meanderings, and vulnerable moments wanting to throw it all in the scrap heap, but eventually a solution came to fruition with IBM.
“It’s been an amazing journey that I’ve had in many decades,” said Sara. “IBM stepped up—they said, “you know we don’t really have this”—but this wonderful engineer overnight introduced a cobbled together amazing thing and said, “here's what we did, I think it's a good start.”
Sara agreed, and together, IBM and Humana collaborated to build a proof of concept project, an early demo of what would become a Humana virtual agent powered by Watson Assistant.
How to round up an “Angsty AI Teen”
“When AI starts it doesn't know anything. It's like a baby. You literally nurture it. Now I think I've got an angsty teen. But to see that maturity and evolution in such a short time period is absolutely remarkable.” – Sara Hines
Codeveloped and coproduced, Humana and IBM recently launched a pilot program involving more than 120 providers and the Watson powered virtual agent: “Hi. Thank you for calling Humana. How may I help you? I am a new conversational virtual agent here to assist you.”
The test was blind, to get a sense for how callers would react when confronted by the AI. This was also useful for working out necessary tweaks and updates to continue evolving the virtual agent with all the nuanced ways providers ask questions.
Quickly, providers were able to ask for things they never have been able to ask before. Whether it was inquiring into patients’ specific benefits for a rapid response or asking whether authorization was required for a medical condition, the virtual agent was able to dive deeper into the provider’s intents.
These insights allowed Humana to gain a 360-degree view of their provider callers and better anticipate and serve their needs quickly. “We've got a level of insight that we've never had before in these interactions,” said Sara Hines. “IBM has been one of the most amazing partners along the journey. Not a vendor, a partner, and I don't use that terminology lightly.”
Flash forward to results following the pilot program. One of Sara’s favorite things to do is listen to the feedback recordings. "You were awesome. I love you," one provider said about their experience chatting with Humana’s virtual agent.
Sara’s AI, once a baby, was now growing up and providing value to callers. “It is absolutely spectacular for someone to have that personal connection,” Sara said. “It’s cool, He’s got a girlfriend!” she commented.
It is a fact of life: AIs grow up so fast.
Explore how you can solve with AI and see benefits like Sara’s with Watson Assistant for Customer Service.