At Mainstay, we’re pretty proud of the behaviorally intelligent solutions we offer to colleges and businesses — but our technology wouldn’t be nearly as effective without a bit of human genius behind it. We may be a little biased, but we’re confident that our team is the best at what they do: Using chatbots to have meaningful conversations that guide people toward a more fulfilling life.
In this series, we’re shining the spotlight on our teams to give you a peek into the daily work, cool factor (yup, we kick it with Gen Z every day), and expertise that Mainstay’s people bring to the table.
As director of artificial intelligence (AI) at Mainstay, Ben Lambert brings nearly two decades of experience to his role — and before joining our team, he worked on some pretty big projects. “I actually helped build the language understanding for Amazon Alexa,” Ben casually slips into conversation. “That was a lot of fun.” In addition to his work at Amazon, Ben also helped to develop Spotify’s voice control feature — so it’s safe to say that Mainstay’s AI team is in highly-experienced, innovative hands.
Ben and the AI team build and maintain Mainstay’s AI model — which means they’re the brains behind our behaviorally intelligent chatbots. Most chatbots can deliver automatic responses to questions, but Mainstay’s bots answer to a higher standard. Ours add a personalized touch to the conversation through unprompted, empathetic nudges and timely check-ins, in addition to accurately answering questions. Our approach is proven to boost engagement and drive positive outcomes, but it takes a spark of human inspiration to always provide the best response for any situation. That’s where Ben and his AI team really shine. They ensure Mainstay’s chatbots are able to understand each user — whether they write formally or use emojis. Doing that well involves a ton of bot training and testing.
“AI is tricky because it gets things wrong sometimes, and that’s ok. That’s why we’re here,” Ben says. “We’re constantly working to make sure our chatbots deliver the best responses every time, and on the rare occasions when they don’t, we use those opportunities to teach the bots how to have conversations that are even better going forward.”
As they ensure chatbots understand and respond appropriately to conversations, Ben and the AI team collaborate with Mainstay’s Engineering teams, one of which is led by Manan Thakkar. Manan and his team help manage the inner workings of Mainstay’s Engagement Platform, including its dashboards, messaging tools, and the data insights it provides. “Basically, we make it easier for partners to connect with their students or employees — and easier for our teams to determine the best ways to deliver partners the actionable information they need,” Manan says. “We’re working to further personalize messages so that conversations continue to meet the needs of each user.” Both Ben and Manan are aware of the power of conversation to spark positive change — and they’re committed to ensuring our partners’ chatbots are the best solution available for engaging students and employees at scale.
Developing the language of Behavioral Intelligence
Students and employees have questions — lots of them. Whether they’re looking for information on financial aid or wondering about the benefits available at their new job, they need instant answers. This search for information is both a key indicator of engagement and an opportunity for you to build trust with students and employees. Every time you provide your people with accurate, empathetic responses that are relevant to their current situation, you build trust with them — even when you use a chatbot to do it. When your students and employees trust you, you can start to form meaningful relationships and encourage them to take steps toward success.
For those encouraging conversations to result in positive outcomes, they need to combine three key ingredients: Accurate information (IQ), empathy (EQ), and contextual relevance (CQ). We call this approach Behavioral Intelligence, and it’s a remarkably effective method for helping people identify and achieve their goals.
Each component of Behavioral Intelligence works together to provide the best response for any situation. IQ is critical to ensuring your people have access to the information they need — but it’s not simply what you say that matters. How you say it makes the biggest difference, whether you’re providing an answer, a friendly check-in, or a timely nudge. That’s where empathy and contextual relevance come into play. Because EQ and CQ are uniquely human traits, it takes a spark of human inspiration to build them into automated conversations — and that’s when Ben and Manan join the party. Through their work building and maintaining Mainstay’s AI and Engagement Platform, they ensure that our bots understand messages from students and employees — and use that understanding to provide behaviorally intelligent responses.
“We cast a wide net on our data to better understand what people need — and what they mean — when they ask a question or respond to a message they receive from our bots,” Ben says. “For example, students might ask about the FAFSA in a range of different ways depending on their circumstances. They might be nonchalant and straightforward, or they might be stressed and desperate. Basically, the messages people send to our bots can be phrased very differently depending on who’s sending them and what’s likely to be going on in their lives. We try to help them out either way, and everywhere in between.” By ensuring Mainstay’s AI model understands a wide range of tones and contexts that a sender may use, Ben and the AI team can make sure the system is capable of providing a behaviorally intelligent response for any situation.
Addressing bias to engage diverse audiences with meaningful conversations
As they consider how our bots will interpret and respond best for each situation, Mainstay’s AI and Engineering teams always keep the audience in mind. Manan is especially aware of the impact of delivering specific groups of students or employees the best information for their unique needs. “Identifying specific audiences gives a fuller picture of engagement and success,” he notes. “We’re providing powerful insights for each partner, so they can better gauge and set their engagement goals.” When partners are able to identify and target customized, contextually relevant messages to specific audiences, they can further tailor ongoing conversations to fit the needs of each group of students or employees. This is crucial for ensuring users have access to the information they need, when they need it, delivered in a tone that will meet their needs and expectations.
Manan also advocates for audience awareness when developing behaviorally intelligent messages. “We want to ensure information is accurate, empathetic, and contextual — and when we determine and develop conversations for a distinct audience, we can make sure the information we provide is relevant to them,” Manan says. He notes that interacting with students and employees from different audiences is an interactive, dynamic process that involves constantly determining what people need — and developing the best message to deliver that information.
While they work to meet the needs of a diverse range of audiences, both the AI and Engineering teams are aware of a concerning issue in AI: Bias in data. “There’s nothing inherently biased about AI,” Ben notes. “However, AI systems are trained on data, and if that data has a degree of bias in it, then the system itself may be impacted. For example, an AI bot can get confused by certain dialects or phrases if it hasn’t encountered examples of that specific dialect before.”
When this happens, it can create a negative experience for the person chatting with the bot. That’s why Mainstay’s teams are extremely aware of the importance of unbiased data, and are working to combat bias in our AI. Ben adds, “Our triage and support teams are progressive and aware of challenges that students and employees face when it comes to implicit bias,” Ben says. “They are great advocates for ensuring that underserved groups have fantastic experiences with our bots. This has resulted in countless changes, like giving our bots the ability to learn a person’s preferred names and using that name in conversations, instead of the name that may be listed in our database by default.”
Manan seconds the importance of reducing bias in data in order to better serve students and employees. “It’s critical that our bots respond with personalized, accurate information that helps make people feel heard,” he says. By using the innovative solutions developed by Mainstay’s AI and Engineering teams, partners can ensure students and employees get the information they need — delivered with familiar, conversational language that removes the language barriers many traditionally resilient student populations often face.
Take engagement to the next level with behaviorally intelligent chatbots built for human conversations
Through the power of accurate, empathetic, and contextually relevant conversations, Mainstay makes it easy to guide your people toward long-term, sustained success. Our AI and Engineering teams are dedicated to developing innovative tools that engage students and employees at scale — so you can deliver the automated, personalized messages they need at just the right moment.
Request an interactive demo today to learn how you can spark meaningful conversations that will support your people throughout their entire journey.
