AWS Founder Spotlight: mpathic | Amazon Web Services

Summary

The video features mpathic, a company that utilizes AI to enhance empathy and understanding in conversations, particularly within life sciences and healthcare. Founded by a licensed clinical psychologist, Grin Lord, mpathic leverages machine learning and natural language processing to provide real-time coaching and feedback aimed at improving trust and empathy in communication. The platform has been especially impactful in healthcare settings, reducing events like readmissions by employing empathic listening. Additionally, mpathic has a strong focus on diversity within its team, emphasizing the involvement of women in AI development. The system's scalability and compatibility with AWS infrastructure underscore its operational efficiency. Ultimately, mpathic aims to infuse empathy across various sectors, including HR, sales, insurance, and clinical trials, reflecting the company's mission to enhance human connection through technology.

Highlights

• Mpathic's AI models offer real-time feedback to improve empathy in conversations. 🤖
• The technology has led to a 48% reduction in hospital readmissions through empathic listening. 🏥
• Mpathic utilizes AWS for reliable and scalable infrastructure, making it easy to integrate with customer systems. ☁️
• The company's leadership and team construction emphasize diversity and intentional inclusivity. 🌍
• Mpathic hopes to empower all sectors with enhanced empathic communication skills through their AI. 🔄

Key Takeaways

• Empathy is at the core of communication, and mpathic is pioneering empathy-driven AI solutions. 🧠
• Empathic listening has significant benefits, reducing hospital readmissions and saving costs. 💰
• AWS infrastructure supports mpathic's scalable and secure operations. 🌐
• The company emphasizes diversity with a female-majority leadership and diverse development team. 🦸‍♀️
• Mpathic's technology is empowering sectors from healthcare to HR with empathy-enhanced communication. 💬

Overview

Mpathic, a company born from the intersection of technology and psychology, is revolutionizing the way empathy is understood and implemented in conversations. Founded by Grin Lord, a clinical psychologist, the enterprise uses machine learning to coach individuals towards better empathic communication. The impact of their technology is evident, particularly in healthcare settings—demonstrating a 48% reduction in hospital readmissions when empathic listening techniques are applied.

Central to mpathic's operational success is its partnership with AWS. This collaboration ensures not only the scalability and reliability of their technological infrastructure but also the seamless integration of mpathic's systems into various customer environments. Their use of AWS also highlights the importance of data security and system efficiency in delivering robust empathic communication solutions.

Diversity and inclusion are at the heart of mpathic's ethos. A female-led company with a diverse team, mpathic brings together vast expertise to create AI products that are both inclusive and effective. This intentional approach is not just ethically significant but also enhances the cultural relevancy and robustness of their AI models. Mpathic's vision extends beyond healthcare, aiming to infuse empathy into industries ranging from HR to sales, setting a new standard for AI-driven communication.

Chapters

• 00:00 - 00:30: Introduction to Empathic AI The chapter 'Introduction to Empathic AI' explores the role of empathy in artificial intelligence, particularly in conversation analytics within Life Sciences and Healthcare sectors. Empathic AI emphasizes understanding through empathy, being crucial for accurate interactions and communication. It leverages AI to enhance trust and empathetic engagement in conversations. The chapter includes insights from a clinical psychologist with 15 years of research experience on how specific words and phrases can foster empathy and trust. A case study involving interactions with individuals in emergency departments post-drunk driving incidents illustrates the application of these principles.
• 00:30 - 01:00: Impact of Empathetic Listening This chapter discusses the significant impact of empathetic listening on individuals involved in accidents. It highlights a study where participants received 15 minutes of empathetic listening, leading to a substantial decrease in their alcohol consumption. The reduction in drinking persisted for 3 years, resulting in a 48% decrease in readmissions. This approach not only benefited the individuals but also led to significant cost savings for the hospital due to decreased readmissions and improved care. Inspired by these results, the hospital considered implementing empathetic listening more broadly.
• 01:00 - 01:30: Training and Challenges in Empathetic Listening The chapter titled 'Training and Challenges in Empathetic Listening' explores the process of training individuals, particularly psychologists, in the art of empathetic listening. The narrator discusses their experience being part of a team responsible for traveling nationwide to conduct training sessions in empathetic listening. The key finding of these sessions was that a mere two-day workshop was insufficient for changing behavior or instilling the reflective practice needed for empathetic listening. The need for ongoing coaching and real-time feedback was emphasized as essential for altering communication behaviors effectively. Furthermore, around 2008, advancements in machine learning and natural language processing were highlighted as potential tools to aid psychologists in this learning process.
• 01:30 - 02:00: Integration of AI in Empathetic Listening The chapter discusses the integration of AI in recognizing and providing feedback on empathetic behaviors in real time. This technology enables instant coaching for professionals, akin to sports training, by detecting approximately 200 different behaviors modeled by psychologists. The AI tool not only identifies these behaviors but also offers corrective suggestions, comparable to an immediate response system in diabetic monitoring scenarios.
• 02:00 - 02:30: Advancements in AI Detection and Feedback The chapter titled 'Advancements in AI Detection and Feedback' discusses the integration of AI in therapeutic settings. AI models, by combining the expertise of hundreds of doctors, have surpassed human abilities in accurately detecting certain conditions. These models can measure the ideal ratio of reflective statements, open-ended questions, and affirmations that contribute to effective therapy, ensuring that patients receive empathetic and precise care.
• 02:30 - 03:00: Empathy’s Role in Healthcare Outcomes The chapter 'Empathy’s Role in Healthcare Outcomes' explores how empathy contributes to improved healthcare outcomes by enhancing rapport and trust between healthcare providers and patients. It cites research on the science of communication and how emotional intelligence is crucial in this field. The chapter also mentions a company that employs emotionally intelligent individuals, strengthening the company's mission and creating a collaborative and empathetic environment. Furthermore, it highlights the understanding of both customer and environmental needs, reflecting the company’s core values and focus.
• 03:00 - 03:30: Client Relationships and Trust Building The chapter 'Client Relationships and Trust Building' discusses the importance of trust and building strong relationships with clients. It highlights the role of empathy in communications and products, emphasizing the need for partners who are both scientifically backed and user-friendly. The chapter also includes a case study of a Seattle-based customer aiming to enhance candidate conversion rates.
• 03:30 - 04:00: AWS and Scalability This chapter discusses the importance of empathy in candidate selection processes. It highlights a metric that combines empathy with other scores like patience and fairness, which impacts candidate conversion rates. With a certain empathy threshold, there's an 8% higher candidate conversion rate. The chapter also emphasizes AWS's role in scalability through its serverless cloud architecture, beneficial in customer relations.
• 04:00 - 04:30: Security and Integration with AWS The chapter discusses the use of AWS services to facilitate enormous scalability in a small startup environment. It highlights the utilization of ECS Fargate for hosting and inference, S3 buckets, containers, and DynamoDB to build and run models in real time. Security is emphasized as a top priority, with a HIPAA-compliant data pipeline ensuring encryption in transit and at rest.
• 04:30 - 05:00: AI in Communication Enhancement The chapter discusses the use of AWS in enhancing communication experiences for businesses. The system's ease of integration with customers already using AWS is emphasized, highlighting the ability to connect systems swiftly via S3 buckets. The reliability, speed, and security of AWS add to its appeal. Additionally, even if a customer isn’t using AWS, their systems can be integrated seamlessly due to AWS's scalable setup. Moreover, the chapter underscores the pursuit of purpose-built AI solutions for particular and highly effective applications.
• 05:00 - 05:30: Diverse Leadership and Impact The chapter explores the complexities of communication, highlighting the nuances such as hand gestures, eye contact, and vocal inflections that humans use to convey meaning. It emphasizes the challenges in understanding these aspects of communication and the potential of AI to bridge these gaps. The discussion underscores the need for expertise and data to create significant impacts, particularly in diverse leadership scenarios.
• 05:30 - 06:00: Investment and Support for Female Founders The chapter discusses the emphasis on female participation and support in the realm of AI and investments. Data collection for models involves women extensively starting from the data annotation stage. Furthermore, there is a notable focus on female leadership within the organization, as evidenced by an entirely female board and investors. The executive team is composed entirely of women or non-binary individuals, highlighting a commitment to diversity and inclusion.
• 06:00 - 06:30: Empathic's Mission and Broader Reach The chapter titled 'Empathic's Mission and Broader Reach' discusses the strategic and intentional choices made by the executive to diversify the team by including women, BIPOC, and other underrepresented groups. The company believes that such inclusive strategies lead to better-performing companies and more robust and generalizable AI products. The executive emphasizes the importance of cultural awareness in product development, advocating for more people to make intentional choices towards building better AI.

AWS Founder Spotlight: mpathic | Amazon Web Services Transcription

• 00:00 - 00:30 [Music] empathic uh produces conversation analytics for Life Sciences Healthcare and Beyond and we focus on AI empathy because all accurate understanding and conversation starts with empathy I'm a licensed clinical psychologist by training and I have been spending the past 15 years of my life researching what are the exact words and phrases that lead to increased trust and empathy in conversations we were talking with folks coming into the emergency department after a drunk driving
• 00:30 - 01:00 accident and providing them 15 minutes of empathic listening and we found that the folks that got the empathic listening had significant drops in their drinking and it held for 3 years they had a 48% reduction in readmission just from being listened to and understood and like radically accepted for who they are in that critical moment and when that happened the hospital saved you know millions of dollars uh care you know increased massively and so they said okay let's bring in all of these
• 01:00 - 01:30 psychologists to listen with empathy someone needs to train them so I was part of that team that actually ended up going around the country training people in empathic listening what we found is that in a two-day workshop you can't actually change your behavior and learn to listen with empathy what people needed was coaching they needed real-time feedback on how to change their communication behavior and around 2008 is the first time where machine learning natural language processing came onto the scene and we said could this be something where psychologist
• 01:30 - 02:00 could listen to and annotate these recordings for empathy for these behaviors and then give automatic feedback to the doctor in real time so we don't have this lag time and they can get coaching just like learning a sport so we have about 200 different behaviors right now that we've modeled with AI from our psychologists that are now being deployed instantly we have a detection of those 200 different behaviors and then a tip and then we actually tell them what to say to correct and get like back on course if you think of things like diabetic monitoring a person is having a tough
• 02:00 - 02:30 time with that therapy and they reach out to somebody for some help with that and they get a empathetic and um open-ended questions and they're with a person who's giving them the help they need they will continue using that therapy RI combines the experience of hundreds of doctors and so we've ended up actually being more accurate at detecting some of these things than humans because of that combined expertise we know exactly the ratio of reflective statements to open-ended questions to affirmations that lead to
• 02:30 - 03:00 improved Healthcare outcomes improved Rapport and trust this is backed by research we know and understand the science of communication and we can model it grin hires really emotionally intelligent people so it does really bolster um the mission behind the company she has an understanding of what needs are of what customers need but also just the environment she's created empathic is so collaborative um and empathetic she really shows the values of our company main focus is on um
• 03:00 - 03:30 building trust and building these relationships with clients and ways that she can help them we've been using the platform for about 2 and 1 half years so as we were thinking about bringing empathy into the candidate conversation and into our product itself it was really important to us to have a partner we could work with that was not just science-backed but it was usable in a way that made sense to our customers so one of our customers um actually here in Seattle was looking to increase candidate conversion so they were having
• 03:30 - 04:00 a lot of candidates that came into their process opt to leave the process so what we found in our metric that combines empathy powered by empathic a few other scores around patience and fairness we found that when that score um was higher and hit a certain threshold there is an 8% um higher rate of candidate conversion which was a big outcome for a customer we were AWS from day one with our serverless Cloud architecture which meant when I would go to a customer that
• 04:00 - 04:30 has let's say 20 million API calls in a day I could instantly as a very small startup scale to to serve the models there we use ECS fargate uh for all of our model hosting and inference as well as S3 buckets and containers and Dynamo DB so we use the full Suite to both build the models and run them in real time with inference security is extremely important to us we have a hippoc compliant uh data pipeline where things are encrypted and Transit and at
• 04:30 - 05:00 rest AWS makes our lives so easy and it's really nice too when we find a customer that's also on the AWS system because we can just connect our buckets on S3 get the system going you know in minutes uh and again we know that it's secure we know that it's fast and everything's working together even when we find a customer that's not on AWS we can integrate them well because of the way that the system is set up for scalability so we've been just really happy since day one we were looking for purpose-built AI very narrow use cases that are very powerful
• 05:00 - 05:30 and that really involve needing an expert with a data set in order to produce something outsized and outstanding and as soon as I talked to grin it started really clicking in that this is big communication is really complicated I use my hands I use my eyes I use words inflection all of those things being read by another human is more challenging than we all believe and one of the things that AI can do is
• 05:30 - 06:00 really enhance that opportunity for a more robust connection so every piece of data that we've collected is annotated by um someone who maybe hasn't had the opportunity to be in AI there's a lot of women building our models that's starting at the level of the data all the way through so every one of my investors is a female check writer our board is 100% women uh all of our Executives they're female or non-b
• 06:00 - 06:30 or our bipac person of color and those are all choices that I made intentionally to build the company and what happens when you do that obviously we have research that shows that companies that do this perform better but we have a better AI product it's more robust and generalizable it has more cultural atement you as an executive can choose to do that I think those are very intentional choices that I made and I you know wish more people would make um to build better AI the thing I am looking
• 06:30 - 07:00 for most to in an investment in empathic is its impact I believe it will have a very large impact and the fact that it brings in this Behavioral Science perspective and and is rooted in data is something that's really important to me most of my advice for women Founders isn't for the founders themselves it's for everyone else we just need more support I think from the community does take that initial investor or in initial support or partner to say of course you
• 07:00 - 07:30 have this but I will say that about female Founders because we do get you know funding I think it's like less than 2% of funding what that means is we have to pitch way more I'm a former founder myself so I became intimately familiar as a woman founder with some of the unique challenges and barriers that women face in the startup space and I decided to move to the investing side of the equation to help to remedy some of the inequities and disparity that exist in Venture Capital so when we came AC cross empathic it was very clear to us
• 07:30 - 08:00 that the CEO GRL Lord had a very deep understanding of the technology really had that perfect combination that we look for of the domain expertise as well as the the technical knowledge to be able to inform those decisions and right now I'm focused on bringing empathy to everyone so we work with uh HR sales insurance and most recently in the Life Sciences in clinical trial monitoring anyone no matter where they are could learn to listen with empathy get feedback on things that maybe they've
• 08:00 - 08:30 never received feedback on self-train and improve their listening so this is just amazing where we're at with technology now that in real time I can get feedback on that in a way that most people wouldn't have access to [Music]