In this episode, Claudine Ryan, President of TalkingPoints, provides a deep dive into TalkingPoints’ mission and the results they’ve seen, how Snowflake has helped them scale, and her team’s ultimate vision of leveraging AI for good.
In this episode, Claudine Ryan, President of TalkingPoints, provides a deep dive into TalkingPoints’ mission and the results they’ve seen, how Snowflake has helped them scale, and her team’s ultimate vision of leveraging AI for good.---
BUILD 2024
Dive deep into the world of apps and generative AI from Snowflake BUILD. Catch up on the latest announcements focused on building apps, data pipelines, and machine learning workflows in the age of LLMs.
Learn more and register for a BUILD event in your region.
[00:00:00] Producer: Hello and welcome to the Data Cloud Podcast. Today's episode features an interview with Claudine Ryan, President of TalkingPoints. In this episode, Claudine provides a deep dive into TalkingPoints’ mission and the results they've seen, how Snowflake has helped TalkingPoints scale, and their team's ultimate vision of leveraging AI for good.
So please enjoy this interview between Claudine Ryan and your host, Steve Hamm.
[00:00:29] Steve Hamm: Claudine, it's great to have you on the podcast.
[00:00:32] Claudine Ryan: Thank you. It's really nice to be here.
[00:00:34] Steve Hamm: TalkingPoints is a mission-driven organization. Tell us why and how Heejae Lim, the founder and CEO, started the company and what you all hope to accomplish.
[00:00:45] Claudine Ryan: Sure. Happy to do that. Heejae founded TalkingPoints based on our own experience growing up as a Korean immigrant student. And seeing her mother make a difference in her education because she had the voice to do so. Because of this, TJ realized a few things. First, that her story wasn't unique, that there are 40 million students here in the U.S. growing up in under resourced and immigrant communities. Secondly, that her experience of family engagement was actually a topic with Lots of research, and that is a very powerful lever that can improve student outcomes. The reality is that family engagement has been proven to be twice as effective as family socioeconomic status in predicting student outcomes, but yet it's the most under resourced lever we have to do that.
If you think about it, it's incredible to know that we have a strategy that can bridge the immense equity gaps that exist in our country. The third is that technology could make family engagement accessible and equitable. The traditional models of family engagement are high touch and not very scalable.
One of the most proven strategies is a homeschool visit, which is fantastic. But imagine if you could do that. Take the benefit of a homeschool visit and make it accessible to all families via technology, things like that. The other piece of it that I want to mention is that teachers receive little to no training in their own professional certification as a teacher, and that's really unfortunate because that makes it really intimidating.
Difficult, time consuming, stressful for them, often, to have the interactions with families that can make such a huge difference to students. So now, with the advent of technology, particularly AI, there are just more, the sky's the limit, I feel like, in what we can do to really leverage family engagement to improve student outcomes.
[00:02:33] Steve Hamm: It seems like your mission has become even more critical since COVID. Is that correct?
[00:02:38] Claudine Ryan: Yes. COVID certainly highlighted the needs for schools to, to, to build strong partnerships with families, particularly the most under resourced families that were hardest to reach. But even, but, Post COVID, it's alarming that the staggering rates of learning loss and chronic absenteeism continue.
Research is showing that kids have lost two decades of learning, and the problem is only getting worse. Plus, nearly 25 percent of students are chronically absent, and mental health and well being issues are a pervasive problem. These challenges disproportionately affect under resourced students, causing even bigger equity backs.
gaps. And family engagement can be that answer to closing those gaps, but historically it's happened far less in under resourced communities because of barriers such as technology access, language, limited time and capacity, etc. In addition, schools are struggling to implement best practices. They need to be embedded as a core strategy through their systems.
They're trying, but current practices aren't really realizing the full potential that family engagement can offer. To make it easy for families and educators to build strong relationships, there are various ways to do that. Attendance and moving the lever on attendance is one of the areas that we've proven through our research that can really benefit from family engagement.
To date, we we are showing through our current Recent quasi experimental study that we've been able to improve attendance, an average of 15 percent for all students, and even more so for the most under resourced students. Similarly, in areas of academics, we have, we have similar results, and these are really fantastic examples of equitable outcomes, outcomes that really.
Uh, bring all students up, but particularly those that are struggling the most. Is
[00:04:19] Steve Hamm: the platform targeted specifically at low income and immigrant communities, or is it being used by all sorts of communities and all sorts of schools?
[00:04:29] Claudine Ryan: That's a good question. We hear consistently that districts are struggling with providing consistent, high level, equitable outcomes.
That means outcomes support all students, all families, where they are. Our universal family engagement platform is designed to drive outcomes for all students. Our target is particularly to reach the 40 million students, that's 4 out of 5 in the U. S., that come from under resourced communities and are likely to fall behind if they don't have that support.
And we're being used in all types of schools across the country. In fact, 10 of the largest school districts use TalkingPoints, and we're in 2 out of 3 schools nationwide.
[00:05:08] Steve Hamm: How does your universal family engagement platform work on a day to day basis?
[00:05:14] Claudine Ryan: So, day to day, educators or family members can reach out to one another.
Let's talk about the families in this case. So a family may reach out to an educator to give an update on their student, to talk about something that they're struggling with, to explain, I don't know why they missed class or homework assignment, that message will be translated from the family's home language to English.
Such that the educator can read and understand and respond in support of that student. And that message is written back in English and then comes back translated into the family's home language. That's kind of the super, super duper basic part until we layer on the different ways to encourage and nudge and support best practice.
Best practice conversations between educators and families and provide resources that support those.
[00:06:06] Steve Hamm: Are the conversations ever live on the fly or are they all text?
[00:06:12] Claudine Ryan: They're SMS, text, or mobile app based messages. Yeah. Although we do offer both video messaging from educators or staff to families that is translated with closed captioning in the family's home language.
And we also offer audio messaging from schools and districts to families. And that is. Translated verbally, we transform a cell phone, a tool that's in the hands of almost everyone these days, into a powerful AI driven education intervention that coaches, explains, translates, and guides families and educators on how to best partner to help children.
We provide actionable data to schools and educators that guide family educator partnerships to share critical information on student progress, provide best practice resources to facilitate stronger student engagement in the classroom and at home, and our impact is proven. Research shows the tachymoid drives student attendance, math, reading, well being, and other learning behaviors.
We recently added document translation as well, so that's been really a wonderful new feature. We're actually releasing it right now that allows educators to translate longer form. Let's say it's like, you know, a weekly newsletter or a permission slip or other content like that that can also, that goes into those same 150 languages to families. So I value that.
[00:07:30] Steve Hamm: Cool, cool. I understand that you've described TalkingPoints as an AI first organization. What do you mean by that?
[00:07:38] Claudine Ryan: We consider ourselves an AI native solution, meaning that our approach has been centered on AI. I would just say more broadly, I'm using technology that helps to solve the problems of our users and partners since we launched in 2015.
And that has very much included AI from the very early days. So we consider ourselves experts in leveraging AI for good. It powers our day to day approach, our product innovation, our impact research. So as an example, AI powers our ability to translate TalkingPoints into over 150 languages. And we are able to do that using custom machine learning models that build on our own human translated content.
As well as a lot of investment in breaking down hard to understand terminology that is specific to the U. S. education system that makes it a lot, a lot just more possible for, for those who don't have that background to, to navigate our U. S. education system. And even we who do speak English fluently, it can have, it can have difficulties there.
So all of that is, is super helpful and, and, and effective.
[00:08:44] Steve Hamm: So, you've been using technology, including AI, from the beginning, from, from, from 2015. But you know, in the past year and a half, we've had this revolution in AI with large language models and, and generative AI and, and chatbots. Are you using any of that stuff?
[00:09:02] Claudine Ryan: We absolutely. Some of the things I didn't mention earlier, actually, are using AI to mine our conversations, understanding top topic, quality, sentiment. Things like that, that have really helped us to, to launch features or to, to, to build on our own understanding and therefore launch features that really help to guide best practice conversation towards improved student outcomes.
We also recently launched a solutions incubator that allows us. to experiment, particularly with AI features and kind of rapid cycle development and learn. So some recent things that we've done are creating ability to have personalized, culturally relevant engagement that's tailored to the unique needs of students and families.
Things for example, such as using AI to create personalized activities for students to do at home with their families based on specific criteria, such as topic area, curriculum standard or learning. So, for example, if a third grade teacher has a student struggling in math, they can fill out a short profile and AI will instantly draft a message to the student's families with a grade level appropriate activity to continue learning math at home, and it can even tailor the output to specific standards, such as Common Core or state academic standards.
Another AI innovation is our is ability to drive in the moment guidance on best practices. We have a new generative AI feature called Message Mentor, and we've trained this model using the dual capacity framework and other research on best practices that we've built over time. And it allows educators to craft messages that are positive, that are asset based, that really reinforce and support building a positive partnership with families, even when they're having to have More difficult conversations about things that a student might be struggling with.
As you mentioned, the benefits are, are, are wonderfully impactful to educators. And we're so excited about what the future holds with the just rapid progress in, in, in LLMs and their ability to, to, to give us more and more features and functionality in AI. The reality is that teachers can't work any harder and burnout is an all time high.
They need effective tools that make their jobs easier and lighten their loads. And it's really important for us to be able to use those to foster trusting relationships between family and school and remove barriers to family engagement, but also guide the nature of those relationships towards behaviors that research shows will have the most impact on student learning.
[00:11:27] Steve Hamm: You've been with TalkingPoints for five years, first as COO and head of product and now as president. What's your strategy for increasing your scale from 4 million students, which you have today, to 7 million by the end of 2026, which is, which is your new target? And what role will technology play in that scaling?
[00:11:48] Claudine Ryan: Yeah. So thinking back in 2020, only a year after I joined, the demand for our product skyrocketed exponentially. We went from a quarter of a million educators and students, students, families to over a million, and from there to meet the demand and drive scale at that time period, we strategically invested in our reach and impact.
We offered our licensed product free to high need schools and districts during the pandemic. Many took us up on that. Many of us have stayed with us for the year since because we've continued to build on our product offering, continue to allow them to support effective family engagements in their schools and districts while still having the free product for individual teachers who want to use TalkingPoints.
So we have a licensed model for schools and districts that they pay for that provides a consistent, equitable system wide. approach and this strategic decision supports our long term sustainability and scalability as a nonprofit. And none of that would be possible without the efficiency and scalability of AI and technology.
Looking to the next two to three years, accelerating to a market tipping point by reaching those 7 million students that you mentioned earlier. Modeling the gold standard. For family engagement to shift mindsets and make effective family engagement a universal practice and building the field and expanding the evidence base for family engagement.
These three pillars work together to create a flywheel of adoption of universal family engagement practices and impact at scale.
[00:13:18] Steve Hamm: Do you see an important role for technology in kind of this next phase? Are there new attributes or new features that you expect to be able to use?
[00:13:28] Claudine Ryan: You know, it's interesting. Our vision began with breaking down the barriers to effective family engagement, translation being a really important part of that. If you can't have a conversation where you really understand each other and trust that you're getting your important information across, that's a really important starting point, right?
But we've always intended, the vision has always been to guide. I wouldn't say copile is necessarily the right word. Maybe coach teachers and families to have those effective conversations. We've really developed a lot of expertise over the years and how to do that. And our vision has always been to embed that into the product.
So the right message at the right time to deliver the most effective results. That's something that happens naturally in our product, has always happened naturally in our product. The best teachers know how to, to do that. I don't know where they find the time in the day to have the perfect conversation about every student with every family, etc.
Those, those happen all the time. But imagine if we can use technology to make that doable, easy, almost seamless for all teachers and families, so they can all have those really productive, effective conversations at the time that they need them. And AI, that vision was always there. AI is just finally at the point that we can really accelerate our progress towards that vision.
[00:14:46] Steve Hamm: So when and why did TalkingPoints start using the Snowflake data cloud? And how do you plan on using it in the future?
[00:14:54] Claudine Ryan: We started using Snowflake at the start of the pandemic when our user base was growing exponentially from a few hundred thousand to more than a million educators and families.
We needed a database that could support our rapid growth and help us scale further down the line. Then and now we have a lot of data. There are over 4 million users on our platform. As you've mentioned, over 500 million messages that have been exchanged to date. And it became more and more difficult for us to use our own homegrown analytics tools.
To support our needs to make smart business decisions and to use our data in all the most effective ways. So Snowflake has helped us to support analytics on the platform, but it also is a source of internal data and analytics that power our business intelligence and research needs. So by that, I mean, there are external users.
and partners, schools and districts around the country have access to analytics that help them make smart decisions about how to use TalkingPoints in their school communities. And then we use it internally for all the purposes that help us to drive smart business decisions around how we invest and make the most of our data.
[00:15:59] Steve Hamm: When families, you know, students and teachers are using the platform, is Snowflake Data Cloud always there kind of engaging and enabling, or is it something that's kind of like done separately for data analytics and, and things like that?
[00:16:15] Claudine Ryan: Snowflake is used within our platform for school and district administrators to understand how TalkingPoints is working.
So for example, messages being sent, responses from families, understanding them by all the different. Breakdowns of, of those families, English-speaking families, English language learners, by language, et cetera. Understanding schools that are highly engaged, teachers that, teachers and families are highly engaged.
Those that might be struggling a bit more and might need more support and guidance. Yeah. Things like that. That's within the product.
[00:16:50] Steve Hamm: So you're doing a lot of data sharing with the administrators. It sounds like.
[00:16:54] Claudine Ryan: Yeah. And then also consulting them on how to take most advantage of their data, because you can present data or you can make it really actionable and useful. And that's, that's an important part of it too.
[00:17:04] Steve Hamm: Right. Now, a minute ago, you talked about your solutions incubator. I want to hear a little bit more about that. I mean, when did you set that up? How does it work? What kind of results have you gotten from it?
[00:17:18] Claudine Ryan: Great question. It's one of those areas that we're super excited about.
We've always invested in innovation in our, in our work, and it's always been kind of like a, I don't know, it's been a little bit here and a little bit there, but this year we were really thrilled to be able to resource a small lean and mean team. That we call the Solutions Incubator that is a cross functional effort between research, data, product, and engineering, and we see it as an innovation hub, a place to rapidly design, test, pilot, and measure the effectiveness of new technologies, features, and services in close partnership with our users.
So, through the Incubator, we run, uh, Many pilots a year with a focus on increasing student outcomes in high impact areas like academics, attendance, behavior, and well being. And this is an area where AI really is very, as in as in many of the areas of our organization, very embedded in the work and particularly now.
So we use research based product innovation to lead measurable. We're focusing on things like reducing absenteeism, increasing enrollment, higher test scores, course completions, and course proficiencies. It's also an area in this incubator where we are getting invaluable feedback from our users, meaning teachers, students, families, and administrators on a, on a very, like almost constant basis to help fuel our, our innovation work.
[00:18:41] Steve Hamm: For your information, there's a lot more to Ogres than people think. You
[00:18:45] Claudine Ryan: really need to dig deep and get to know the real you. And the real up close and personal.
[00:18:51] Steve Hamm: Hladin, we're coming to the end of the podcast. And we typically end on a lighter, more personal note. And I understand that your parents immigrated to the United States from Germany and you were born here, but you only spoke German for your first five years.
What kind of challenges did you face adapting to school here? And how could the TalkingPoints platform have helped you?
[00:19:15] Claudine Ryan: So similar to many of the families that use TalkingPoints, I did grow up in an immigrant family and I was an English language learner in my early years. So my first language is German, as you noted.
And while my parents also spoke some English, they chose to speak English. To have me speak only German so that it would be kind of locked in for life, which it still is. They spoke some English, but they had limited understanding of the U. S. education system and their role in supporting me. I'm a first generation college student in my family.
So ultimately I was successful. However, I can only imagine what a difference TalkingPoints would have made to my educational experience. I think my parents would have wanted to be involved, but never really knew how. I guess overall, when I think about it, um, and I think about my experience talking voice, I, I recognize regularly because we have so much one connection to our users.
The parents care deeply about their children. They care deeply about their children, even if they aren't able to pick them up from school every day or, you know, Go to that, um, you know, back to school night or other school event where you have to be in person because some of them are working multiple jobs and it's just not possible at all, but they really do care deeply.
And even after five years of TalkingPoints, it still feels almost magical to me, the trusted connections that we're able to establish between educators and families. that break through those barriers. Barriers like being able to be there physically and present. And those are transformative to students.
And it's really fantastic to feel like I'm part of an organization that is affecting that change.
[00:20:50] Steve Hamm: Well, it's been inspiring to talk to you today. I live in New Haven, Connecticut, which is a city of 135, 000 people and 40 percent are non native English speakers. And many of those Many of the kids are undocumented so that there are lots of hurdles and lots of barriers to success.
And it's amazing to see TalkingPoints and see how useful that could be here. So if you want an introduction, I'll make it.
[00:21:18] Claudine Ryan: Oh yeah. Yeah.
[00:21:21] Steve Hamm: The other thing is, I mean, I just think, you know, there's so much excitement about AI right now, and it's really easy for me and the people at Snowflake to see how this will be useful for corporations to make them more responsive, make them more efficient, you know, more successful in all sorts of ways.
Ways, but talking to you, it really becomes clear how this technology can help regular people. And I think this is an area where I'd say we've, you know, it's underexplored in our society. And I hope that, that. Many, many, many organizations like yours, but in different domains, start to look at AI and start to look at these new technologies and figure out how they can be used for regular folks.
So, uh, thank you so much for being on the podcast today. It's been great.
[00:22:00] Claudine Ryan: Absolutely. Thank you so much.
[00:22:12] Producer: Focus on building apps, data pipelines, and machine learning workflows in the age of LLMs. Watch now at snowflake. com slash build.