In the third episode of Moving Digital Health, the new podcast from MindSea, CEO Reuben Hall talks with Mahshid Yassaei of Tali.ai. Mahshid is a co-founder and the CEO of Tali, one of two digital healthcare companies she has founded in less than 10 years. She and her team at Tali have developed a cutting-edge software based on artificial intelligence (AI) and natural language processing (NLP) that stands to benefit thousands of physicians and their patients.
Mahshid discusses her background in computer science, which was focused primarily on cryptography and data security, and how she and her co-founder made the decision to dedicate their combined expertise exclusively to healthcare (making Tali a true rarity). She elaborates on the stages of the company’s development: how answering a challenge from the Public Health Agency of Canada (PHAC) led to the founding of Tali.ai, how the COVID-19 pandemic created a pilot opportunity, and how the Tali team is approaching their game-changing work.
Much of the information physicians work with falls into the category of unstructured content, Mahshid explains. This poses a problem for prompt information retrieval, but the Tali team has a solution in the works. Mahshid explains how Tali proposes to save time for physicians at the point of care and how beneficial that can be for doctors and their patients alike. She delves into some potential use cases for technology like Tali and also expands on the challenges inherent in working on cutting-edge technology.
As an entrepreneur, Mahshid has transitioned from software developer to CEO and leader. She shares the learning processes she found effective as she progressed along this path, and she discusses the unique challenges she faced as a woman and a non-native English speaker in the tech space. She offers insight on how we might make our industry more inclusive and more inviting and in so doing, hopefully inspire more women to consider careers in tech.
Mahshid also shares her favorite sources for inspiration and learning in the health tech industry, as well as the exciting revolution she sees emerging in the digital health space in the next few years. Mahshid’s experience as a two-time business founder and as a software developer who became her own CEO lend fresh perspective to this discussion of digital healthcare. We thank her for sharing her story and her insight with us, and we hope you’ll enjoy her episode of Moving Digital Health.
Listen to Moving Digital Health on Spotify and Apple Podcast!
Please subscribe to the MindSea newsletter so you will be notified about future episodes!
Read Transcript:
Reuben (00:00.04)
Welcome to the MindSea Podcast series. Moving Digital Health. Our guest today is Mahshid Yassaei, founder and CEO of Tali.ai. Maybe if you could just briefly describe your background to get started.
Mahshid Yassaei (00:00.21)
Sure. Yeah, I did my masters in Computer Science at McGill. Close to about ten years ago. I worked for a BlackBerry, then a few other companies. And then in 2015, with my co-founder, we started to even set up, which was my first entrepreneurial journey and even said was a software company specialize in building solutions for health care and medical industry.
So over the years, we work with a wide range of organizations, from small startups to multibillion dollar organizations to government agencies, and we help them build their solutions and take it to the market while being compliant with the solutions that were the regulations that apply to their products. And then in 2019, Public Health Agency of Canada posted the challenge on their website which was how the tallies idea and the problem the pain point that we identified was basically initiated.
Reuben (00:01.41)
Cool. So even like, you know, early on you got into the health care industry, was there any kind of reason or anything that brought you in that direction?
Mahshid Yassaei (00:01.55)
Well, my background in computer science was mostly on the cryptography and data security. And I co-founder’s was also on the biotechnology. And when we were thinking about, like my role was a security software developer. So every organization I was going to, I was responsible for building solution that deals with sensitive information of users to obvious use cases for this kind of expertise is a financial market and healthcare.
And when my co-founder and I were thinking, we realized that this specialization in health care is specifically meaningful because of this reason, and also that in healthcare there is always a gap between people on the technology side and people that are deep into the health care system. Every side is somewhat intimidated by the other side’s expertise. And we we thought that if you build a team that specialize and have experienced years of experience building solutions for this market, this accumulated experience can be super helpful from the product design market research to product development, identifying all common risks to testing and validation and all that.
And this proved to be the case over the years as we move forward. Every new client was very excited to hear that health care is all we do and we are focused and passionate about what’s happening in health care. So for that reason, I guess that ended up being a good many to focus on.
Reuben (00:03.41)
Yeah, very insightful. It always helps to have the niche. So tell me about the early days of tablet. What does it mean to look like when you’re first getting started?
Mahshid Yassaei (00:03.55)
Yeah, we were kind of lucky with how Tali was started. Like as I said, them, it was to a challenge. The high school said they, the, the problem they had was that they, they spent a lot of time and resources developing evidence based resources for providers of the point of care. But they were not seeing the type of engagement they were looking for because it was very time consuming for providers to find the answer to within the specific questions they had.
One example was that the PAC spent a lot of time developing and keeping the Canadian Immunization Guide up to date. So this guide is a five chapter book, and it has information about everything around immunization and vaccination that IT provider needs in Canada. But it’s because it’s a long guide. A lot of times providers would just Google do questions or will look to American resources just because it’s faster to find the answer that way.
So the solution that we proposed to them was a question answer interface where providers would just ask their questions in natural language and get the answer back from their resources. So we did their preliminary research with P&G on a state of the art MLP algorithms, and we built a portal type and we showed some promising results. So we did even a pilot on COVID 19 content in Alberta.
And then when we realized that, yes, this is a pain point for physicians, they are spending a lot of time digging into all sorts of on a structured content of the point of care. That was when we would you certainly as a separate business and I move to the full time right now, even so this being led by a managing director, that’s amazing and that’s doing a great job there, too.
Reuben (00:06.04)
Awesome. So tell me about the current state of health. What are the different challenges? How is it evolved since then?
Mahshid Yassaei (00:06.15)
Yeah, we did a lot of research discovery on use cases, other use cases for the language model that we’ve built. So this language model understands on a structured medical content and we looked for use cases in the medical space where we are dealing with a lot of unstructured content and we need a fast access, a fast information retrieval engine to that content.
And the the use case that we have ended up with right now is information retrieval from electronic medical records, softwares, EMR So Utah’s and we have built a partnership with well Health, which is the largest the operator of outpatient clinics in Canada, to integrate Tali with Oscar pool and provide physicians with a layer on top of their EMR when they can ask questions like, Is there a history of diabetes in this patient’s family?
And you might find the records from like five years ago or six years ago where a doctor has put a note, where like it says, like this patient’s mother is dealing with diabetes. And if it was not for and solution, they tell you they would have to go to all these records. Many years of notes structured to takes.
Also the structure it takes to find this specific piece of information. But we are basically giving them this layer. There are other features as well. They can use voice to scribe their notes into the EMR directly to ask medical questions like Does citalopram interact with acid? Nelson or what dosage of this medication should I prescribe for these patients, etc.?
So they it’s basically an intelligence layer on top of an MRA, the primary care set out to save time for physicians and basically let them spend more face time with their patients and visit more patients. At the end of the day.
Reuben (00:08.28)
Yeah, it sounds quite simple on the surface, but you know, underneath very complex brain, like to me I’m just thinking of like, okay, well integrating with an EMR, like, you know, technically that’s, that’s very difficult. And then natural language processing has all of its other challenges. So what are the some of the technical hurdles to overcome to connect all that together?
Mahshid Yassaei (00:08.56)
Yeah, you’re very right About all the challenges you mentioned on the technology side, the feed them question and saying from unstructured medical content on a structured content is basically one of the two cutting edge areas of research in NLP. Even if you talk to computer scientists that are working on that side. So it’s not like a technology that’s all over the place is not like a mobile app or web app who are, we know more or less what we’re dealing with or even an Iot device that they think pretty much all where the days where we’re dealing with the technical challenges are still with you In most areas of AI, specifically NLP, specifically in question
and saying AI structured content and we’re still dealing with where do we draw the line, the accuracy line, and how do we deal with cases where the algorithm is not that accurate, especially in a medical case? This is very important because of the liabilities that are there. So there are a lot of challenges on the product design side of it to communicate the accuracy of the algorithm.
How if how confident is the algorithm about the answer it’s suggesting, but it is a game that has both sides like even if you can save an hour or half an hour of physician time, that’s of great value. Obviously with time passing, we will have better algorithms, the more accurate algorithms, and we can save more of their time.
But even we realize that even with the accuracy that we can deliver today, physicians that are still interested are still interested in having that kind of support. So that’s on the technical side. But also, as you mentioned, healthcare is not the easiest symptoms of integrating with solutions and getting to the markets and all the challenges that are there.
But I guess we were lucky to find a great partner like well that were aligned with our vision that this is a piece of technology that needs to be there, that is the future, and they had the means to support this to warehouse, who now has more than, I think, 15,000 physicians in their network. And that will give us a lot of opportunity to develop Tali and perfect for our initial customers while we’re trying to expand to the rest of the market.
So I guess on that side, we’re just I should just say that we were lucky to find a great partner like that.
Reuben (00:11.54)
Yeah. And what does the team look like right now? Obviously lots of specialists and niche areas to to solve all of these problems.
Mahshid Yassaei (00:12.04)
Yeah, we are currently a team of eight people. We have a few NLP scientists in the team, software developers, but also product designers. We are at the see this stage right now trying to show and validate the time that’s being saved in World Health Network and then expand to two other primary care networks.
Reuben (00:12.35)
Interesting. How do you see Tali evolving in the future over the next year or so?
Mahshid Yassaei (00:12.48)
Well, there are a lot of use cases for technology like Tali right now we are our focus in the next year is to perfect these for physicians and increasing it for physicians time and comfort. But Tali can help and support patients as well in between visits. So especially with chronic conditions that a lot of time patients have a lot of questions about the condition, about the medication they’re seeking, and they don’t want to wait till the next time they see their doctor.
So you see a lot of patients just Googling the question and they’re not always armed with the knowledge on how to validate if the if resources evidence based on trustworthy or not. Whereas we can partner with clinics or with the physicians to provide an interface to Tali where the patients can ask the questions and get the answers from resources that are already vetted by the physician.
And the physician will then be sync on what kind of questions, what kind of curiosities they have, and if there is an urgent issue, they can even follow up on that. So there’s so many cases and a is just that is just a long journey and we need to walk it step by step. But it’s it’s a very exciting journey for me.
Reuben (00:14.22)
Yeah. And what about voice? Obviously, when you’re talking about conversational wise, I jump to think, well, you know, the voice UI is a natural progression from that as well.
Mahshid Yassaei (00:14.35)
Yeah, we are already voice enabling. So there is all the interfaces are available both to the physicians typing the questions or talking to Tali. There are some product design challenges around that as well. We imagine that not a lot of physicians are comfortable asking the questions aloud when the patient is in the room. Sometimes they might be or they might want to ask the question or like talk to Tali while they’re walking in the hallways from room to room, just checking something quickly.
So we’re giving them both interfaces, actually.
Reuben (00:15.13)
Amazing. I’m I assumed you had obviously thought about that and come across that you use case. Maybe if I could switch gears a little bit. As a woman entrepreneur, have you ever felt there were additional barriers for you to overcome, you know, during your journey.
00;15;42;24 – 00;15;48;19
Mahshid Yassaei (00:15.42)
Well I haven’t been a male entrepreneur before? So I don’t know about that.
Reuben (00:15.48)
Being an entrepreneur is hard. It is hard enough yet to break through and have a successful product. Yeah, so it was and I’m just wondering if you did ever feel that there is other challenges breaking through, you know, pitching a product or developing those partnerships you needed.
00;16;14;25 – 00;16;42;00
Mahshid Yassaei (00:16.14)
Yeah, that well, I do acknowledge that the game is not equal for men and women and it’s the game is not equal for immigrants and non-immigrants and the native speakers and non-native speakers and all that. So I have a baggage of all these issues. Issues is like meeting you woman and being an immigrant and all that.
So it’s but they have been lucky. I have been surrounded by people that we’re always happy to help and support me and I can’t really complain although I do notice specifically in the B2B markets where or when you’re raising your fund, basically any areas where you were in is subjective opinion of the buyers or the investors. Mother Everybody is obviously subject to some discriminations.
They’re mostly unconscious, I would say I don’t I don’t think most of the time they are. They’re conscious or intentional, but we are all we all have this unconscious biases in our minds. And I think that do have a does have an effect on how how difficult you raise your funds or how difficult you can win a business partner, because just because the other side might not feel confident in trusting you or trusting the quality of the work you’re doing as much as they do the work of another person.
Reuben (00:17.59)
Yeah. And it’s, you know, for my parts, we’re often hiring new people at my and see and frequently the talent pool, there’s just not as many female candidates as there is male for computer scientists. And how do we get more women to pursue career careers in and have that confidence to go after it?
Mahshid Yassaei (00:18.27)
Yeah, I do share the same pain. I do see that there are not enough women computer scientists, and I think the ultimate way is through providing a safe space, an inclusive, inclusive space in in tech. I remember that when I first joined my first job after school, I was a little shock back then. I feel like I didn’t belong because everybody in our team were like this kind of stereotype.
People picture of a cool guy with like your hoodie and the like. You somebody that plays video games, etc.. And when you’re younger and you are at the beginning of your career, you might feel like a stranger to a space like that If you don’t play with your games or if you don’t act or feel or behave like the rest of the team.
So I think diversity you like being inclusive to diversity is the key or has been the key for me. But I think it is a pity that we sometimes see a lot of girls being intimidated by the tech space and the computer science in pursuing that in universities. I hope that changes soon.
Reuben (00:19.58)
Yeah, I really hope that we’re on the right track to bring more equality into the work space and in tech for sure.
Mahshid Yassaei (00:20.07)
What the strategies do you think to hire more women at your own company?
Reuben (00:20.13)
Um, yeah, definitely. I’m happy to talk about that. It is. It is a challenge. Oftentimes we’re going through and posting a job to LinkedIn and we get the candidates to apply to the job, and that’s the pool we have to select with. And we base that pool on the merits and don’t really pay attention to gender. But when that pool is mainly male here, a lot of the times you’re going to end up with a male candidate.
One thing we have done is try to make the job descriptions less specific because some research has shown that females are less likely to apply to a job, but they’re not technically qualified for or many males. We’ll see a list of requirements and qualifications and say, Hey, that’s not quite me, but I’m going to apply anyway. Females are more likely to say, Oh, I don’t qualify for that.
I’ll look for a job that I, you know, check the boxes for. So in that sense, we tried to craft our job descriptions to just make the job seem more approachable and less specific in terms of, you know, exact years of experience and also trying not to ask for the world or a unicorn for the questions. Well, I’ll just try and, you know, try and make it realistic for you and, you know, talk about the opportunities for learning and development and what it’s like to work at mine and see not just who we’re looking for in a dream candidate.
Mahshid Yassaei (00:22.10)
Yeah, that’s amazing.
Reuben (00:22.14)
We also work with a great partner that builds a diversity hiring platform. We’ve been trying to learn from them as much as possible as well along the way.
Mahshid Yassaei (00:22.29)
What are they called?
Reuben (00:22.31)
It’s Mathison.io
Mahshid Yassaei (00:22.33)
Okay.
Reuben (00:22.35)
So Mahshid outside your work at Intel or where do you go to find inspiration or learn about innovation in the health tech industry?
Mahshid Yassaei (00.22.51)
Um, I guess my main source of inspiration is just I follow industry leaders on LinkedIn and Twitter and they I get a lot of inspiration there. They sometimes share their thoughts or share the interesting articles they see and that has been working pretty well. But aside from that, I have been joining conferences, mainly U.S. and Canada, to get updates on the most recent activities in the market.
One example is the CB Insights, the future of health that happens every year. That has been great for the past two years. They have been running that with Chile, but there’s still an amazing gathering of market leaders in Toronto locally. We had the this community event called Health to show that they used to be very active in.
I think there they have paused their activities or even said during the call they hope they come back circle with.
Reuben (00:24.03)
Oh yeah, we’re all looking forward to the post-COVID conference life. I think not that I went to a lot of conferences before, but not having been able to go to a real in-person conference over the last few years. Definitely looking forward to the day we can just kind of, you know, go and be in the same room as people and have those kind of chats, conversations and learn from people in person that we’re getting back to tell you about.
A.I. What are the main challenges you’re expected to face in the next year?
Mahshid Yassaei (00:24.52)
I think right now our focus is to build a product that a few physicians will love and will advocate for and that is what you are focusing on. A lot of the risk comes from building the AI engine, the NLP engine that’s smart enough that they can provide the kind of support they need. And the other is getting a deep understanding of the real pain point at the point of care, getting a clear picture.
How does the can workflow look like? Because we don’t want to. We don’t want to be a burden on physicians, the schedules that are already pretty busy. So I guess it’s some I want to use the word outsider a little conservatively, because we had we do have physicians in their cofounding team, etc. but from, from the point of view of somebody that is not active in the point end point of care, our challenge is to really get a deep understanding of how many different ways a patient who is it goes.
We know some physicians will. They usually call it pajama times. They keep their administrative tasks to when they go home and they start just going to all their administrative tasks, putting things into their EMR, etc.. Some other physicians will take the time during the day between visits, some other do it while the patient is in the room. So really understanding all these use cases and building your product that fits perfectly for what they’re looking for, I think is the point, the main point of focus for us scaling and selling this simple thing and all that comes after that when we have that perfect solution.
Reuben (00:27.03)
Mm hmm. Well, I’m actually wondering what your day looks like as a computer scientist. You spend a lot of time developing and writing software code, but leading a company, your your field looks very different. So what does your day to day look like now?
Mahshid Yassaei (00:27.22)
Yeah, that is actually a huge regret for me. I loved my days when I had the time to develop code and write good software. That was very, very pleasant for me as being a business owner, which has been my life for the past seven years now it’s a different the schedule day to day, every day is a new challenge.
You need to work with a lot of different stakeholders and we lead the vision of the company and be responsible for having enough money in the bank for the rest of the company to operate and in different, different sorts of challenges, which is interesting. If I’m facing new challenges, I have to learn new concepts, new ideas every day, and I love that.
But they also do miss the days when I had time to focus for two or 3 hours on the specific technical challenge and write good code and all that.
Reuben (00:28.31)
Yeah, yeah. I imagine my background is actually in UX and design and I feel the same way that I don’t do design anymore like they used to. But certainly I love to stick my opinion in there. When other people are doing designs. I also think it’s like climbing up and down the ladder some time. As a business leader, you have to have the high level 10,000 foot view and you’re thinking about the vision and the direction of the company.
And then in the next meeting you might be getting really into the weeds on some specific problems, whether it’s on the finance side or the technical side, or are you just kind of jumping all over the place. It’s tough not to be reactive. I always think I wish I could spend more time forward thinking and planning ahead and then, you know, seven things come up during the day that you have to deal with.
Mahshid Yassaei (00:29.42)
Yeah, but that’s the benefit of being a practitioner growing to a business leader because they think, as you mentioned, you, you understand your business to the deepest level and up and that is that is a benefit that not all business owners have. So I guess that’s the positive side of it then.
Reuben (00:30.03)
Yes, for sure. And how’s the learning curve been for you? Because again, coming out of computer science school, they don’t teach you a lot about how to run a business. Right.
It has been very ad hoc and I learned a lot during the process. Just the incubation incubators that we were part of. I use every chance to take the courses that were that were being offered in terms of the marketing side, the sales process, the operations side and all that. And also we have been blessed to get support from great mentors and advisors that were there for us when we had the specific challenges that we need to deal with.
And I have learned so much from the advisors and mentors that we had. That is the that has been really to the point and very timely, The, the, the learning and the education is kind of like a small, longer term investment that comes in. So a combination of both has been very helpful for us to make this transition from a fully technical role to a business ownership kind of so.
Reuben (00:31.31)
Well, you’re doing amazing. Tali sounds like a great company and doing some really cool things. Did you have any questions for me while we’re here?
Mahshid Yassaei (00:31.44)
I saw that MindSea does a lot of developments in the software development. What is your interest in health care? What are you trying to focus or what other industries you focus? I’m curious to know more about that.
Reuben (00:32.00)
Yeah, definitely. Digital health and health is our focus going forward. We’ve been around for quite a while with 2007 is when we first started, you know, when the first iPhone in the App Store came out. We’re part of that. And over the years we’ve developed lots of different apps for different businesses and industries. But several years back we looked and said, you know, what’s the work?
We really enjoy doing that people are invested in and feel rewarded by? And it’s the work we had done with our partners in health and wellness and fitness that really struck a chord and kind of aligned with our values. So over the last three or four years, we’ve really been slowly pivoting in that direction, just focusing on, you know, those partnerships with health and wellness clients and we’re actually working on a complete redesign of the website right now where all of the messaging is going to be focused on on digital health and increasing our specialization in that industry.
So every partner we work with, every person we hire, we’re looking like how we bring new skills and knowledge onto the team, to fill our gaps and really, really own and understand the space. Because, you know, we’re you know, we’re not doctors, we’re not medical professionals, but we really know technology both on the mobile side and, you know, in web and back end.
So we look for that expertise on the medical side with the partners we work with. But the more we bring on new team members that have worked in health tech before at previous positions or do training with the team about, you know, privacy regulations and data security, we continue to level up our specialization in that department Mahshid. I read an article about two women of influence on care and there was a quote I really liked.
It said, I love where I’m at. At the intersection of tech and health care, building more and more tactful products in this space. So I’m just wondering, what are you most excited about in digital space, health space overall in the next few years?
Mahshid Yassaei (00:34.42)
Um, I, I’m very excited about the shift to a health care system where patients have more agency. Um, I, I love the line of products that are being built around that, that basically empower patients through patient education and engagement to remote monitoring devices even older which will cure solutions that are basically there to deliver the care to where the patient is, not where the provider is, which is our care and health care system.
So I think that it leads to a healthier population. And overall, and for that reason, it is a very exciting revolution in the health care system. To me, in and all the different technologies that are being feels to, as I said, empower patients to rise up to this new challenge. A very exciting that we have.
I know a lot of startups that are working on the remote monitoring side and basically devices that measure some metrics from the body could be the EKG signal, it could be the blood pressure, whether body temperature. And then we’ll analyze those measurements to give you updates and and help you with early diagnosis, which we know is very effective in preventing more serious issues and chronic conditions to developing being more serious.
So, yeah, that is what excites me about the health care future.
Reuben (00:36.26)
Yeah, definitely the movement of technology into the home to empower patients to take control and have more agency over their own own care is definitely exciting to me as well. So thank you so much Mahshid for joining us today. And for anyone listening, if you like this episode, please subscribe to my ESI newsletter. Be notified about future episodes.
Thank you so much.
Mahshid Yassaei (00:36.57)
Thank you so much. Have a good day.
