Anurag Kejriwal, Founder at Sonnet AI. Sonnet AI, an advanced AI-powered meeting assistant & notetaker designed to keep conversations in context—helping professionals extract meaningful insights and manage relationships effortlessly in their specific domain– with custom text and visual notes.

As a Computer Science & Engineering student at Ohio State University, Anurag is building an AI platform that goes beyond traditional note-taking, offering a domain-expert, no-bot-needed assistant that understands context and structure. Unlike other AI meeting tools, Sonnet AI doesn’t just capture words—it provides speaker insights, smart summaries, and an inbuilt relationship tracker, ensuring every conversation remains in context.

With features like participant reports and conversation matching, Sonnet helps users keep track of critical discussions across teams, clients, and leads—all in one place.

Listen to the episode on Spotify, Apple Podcast, Podcast addicts, Castbox. You can also watch this episode on YouTube.

In this episode of the AI Minds Podcast, Anurag Kejriwal, founder of Sonnet AI and a full-time student, shares his journey from drop shipping to building an AI startup while juggling college life.

Anurag dives into how Sonnet AI evolved from a sales tool into an audio-first platform that helps legal and startup teams turn conversations into relationship intelligence—with zero bots involved.

He explains how Sonnet's customizable templates cater to professionals across roles, from recruiters to founders, enabling them to extract deeper context from meetings without sacrificing privacy.

The conversation also touches on the technical hurdles of working with audio data and Anurag’s commitment to creating a seamless, bot-free experience—a vision made possible with support from Deepgram.

Anurag reflects on the challenges of growing a company while still in school, his passion for product development, and how AI can personalize and humanize the way we connect through meetings.

Throughout the episode, he highlights the importance of adaptability, continuous learning, and building tools that bring empathy and intelligence to everyday conversations.

Show Notes:

00:00 Pre-Covid Dropshipping Pursuits

04:00 Balancing College, Startup, and Life

08:51 Enhanced Relationship Tracker AI

09:56 Focused Tool for Founders & Recruiters

14:20 "Bot-Free Recording Feature Highlights"

17:14 Balancing AI in Coding

More Quotes from Anurag:

Demetrios:

Welcome back to the AI Minds Podcast. This is a podcast where we explore the companies of tomorrow being built AI first. I'm your host, Demetrios. And this episode, like every episode, is brought to you by Deepgram, the number one text to speech and speech to text API on the Internet today, Trusted by the world's top conversational AI leaders, startups and enterprises, some of which you may have heard of, including Twilio, NASA, Citibank and Spotify. I have the pleasure of talking to Anurag, the founder of Sonnet AI, and this is going to be a very special episode. I think this is our first where we have a founder who is still actually a full time student at university. So it's great to have you here.

Anurag Kejriwal:

Thank you so much for the opportunity, Demetrios. Really appreciate it. But no, it's been interesting, to say the least, being a student founder.

Demetrios:

So I want to dig into all of that, especially around the inspiration behind the tool. But first, let's set the scene. You were born outside the US in India. You decided to migrate, or I imagine your parents decided to migrate and you just were part of that train that came.

Anurag Kejriwal:

Exactly.

Demetrios:

And you now are in Ohio at Ohio State. This was not your first foray into being an entrepreneur. What was your first test and. Or hustle, I should say.

Anurag Kejriwal:

I've had quite a bit. What's really interesting is that like even pre Covid, 2018, 2019, I was really into drop shipping and I used to watch like a lot of YouTubers. But it wasn't just like a normal, like everyday kind of thing. I was really into it to the point where I had like an uber low setup, a Shopify setup, and all these different websites coming together. Ultimately it didn't, end up doing anything because I didn't have money. And I, reached out to thousands of like Instagram theme pages just being can you feature this type of thing? But, I realized that was really fun to me, even though I made zero money on that. So kind of segued from that and then watched a lot of Shark Tank, I'm sure a lot of people do, but it was inspiring to just like see other people get on there and building their own stuff.

Anurag Kejriwal:

Took a little bit of a break from like standard entrepreneurial stuff in high school. there's a lot of different things. I started a STEM club at my high school. But again, not really for monetary value. But it's been a journey. Got into college and realized that I definitely do want to do an entrepreneurial pursuit, whether it is more in software or something else. Because I was kind of late to the.

Anurag Kejriwal:

So I'm a CS and math major and I was kind of late to the computer science trend. I really didn't know what to do with my life for a while. I realized I wanted to do CS like late senior year after I did like an AI camp. So it was definitely like unorthodox getting cs.

Demetrios:

So I've been through college and I will say that I understand how you can do both a company and college at the same time because the classes are incredible. You get to do whatever you want for a little while depending on your major I guess and depending on which school you're going to. But it is much different than high school. How have you been balancing the two as you're trying new focus on launching a startup and also learn more technical skills and all of that fun stuff?

Anurag Kejriwal:

To be honest, I haven't know balance been balancing it to say the least. I'm not the most structured person so sometimes I'm late to kind of facing the problem. Like I just, as soon as I get it, I just kind of sometimes have to wing it. So while I'm doing college and the startup and a couple other things, doing a part time internship as well, I kind of realized the most important like thing throughout all this is more like mind space instead of just even like time for me. I'm constantly thinking about like, the startup when I'm doing something not startup related at all. And I realized like that's kind of where I want to be in terms of like, really focused on the startup. So prioritization has been a huge deal for me and sacrificing some other things that I really like to do the startup. I feel like in college, it is like literally something new pops up every second of the day type of thing.

Anurag Kejriwal:

So it's been interesting but I think I've been able to kind of face stuff head on as they come, but definitely trying to build a structure for the longer term.

Demetrios:

What was the inspiration behind Sonnet AI and when did you formally start it? Break down the story of getting into it?

Anurag Kejriwal:

So I wasn't the first code writer on Sonnet. There were previous co founders who had like a initial idea and app, but we realized, we wanted to take it further than that. So me and my dad actually started working on it. My dad was more of just like advisory role and he helped me out with a lot of things. But being technical, kind of revamped the product, which me and my developer that I also have did. So we were able to develop and revamp the product and, launch it for a different target audience. Actually got into the techstar Spring Columbus cohort and I'm in that right now.

Demetrios:

What was the original product aimed at and how did you change it?

Anurag Kejriwal:

It was very sales focused and less so. I'll get into the company as well, but it was a little less like audio focused. Very like sales reach outs for like B2B companies. And now we're transitioning more to the legal side and also startup teams. So founders, we've been seeing a lot of organic growth throughout them pretty much that was the initial idea, really trying to, you know, see where our organic growth is getting to us. We haven't spent any money on advertising, but we've been seeing a lot of founders and startup teams emailing us, really loving the product. Some have, differing opinions on features. But it's honestly been really cool to get user feedback because I feel like in college, you're not really getting user feedback.

Anurag Kejriwal:

You're building a project for a class, calling it a day type of thing. Right. Which is not a bad thing. But it's definitely a different perspective. And I feel like for me, I learned the best by actually doing something and getting feedback on it. Sounds very cliche, but I think in a startup that's literally, the best advantage you have. You can like hear something today from someone and build it like the next minute instead of having to go through a rigorous process.

Anurag Kejriwal:

That being said, I'm not the best developer. Still learning a lot of things. So I would say the idea is to transition has been really. Because even through my own experience being a student, you're not just a student. You have these five different clubs, then you have these internships. There's a lot of different things you're managing on your plate. And because of that, I've had so many conversations and meetings and I've taken notes on them And it's been great having these notes. But I realized that once I have more than 10 notes, I don't understand the context behind my Conversations, even from just like a summary.

Anurag Kejriwal:

So for humans, I think just through our conversations, we learn a lot from relationships. And through conversations, you're building relationships, understanding the person more and more. And that's like the idea behind Sanne. And what we're trying to do is really build relationships from conversations. So more than just generic, summaries and notes, we want to be able to tie conversations you've had to a person so that you're able to understand the who, what, why, more than just what did I talk about with this person at this standalone meeting? So, going through the founder journey myself has been.

Demetrios:

What does that mean in practice? It's because I, I buy into that vision of I want to go deeper and I don't want just a summary. I want to know, if this is someone that I meet with weekly because I work with them and my relationship is X versus this is a mentor, I meet with them once a month or this is the first time I've ever met with them.

Anurag Kejriwal:

So tangibly that means more that we have these kind of two separate big sections of the app. So one is just your conversation. So your regular notes from in person or virtual whatever conversations you've had. But one interesting piece of the relationship tracker. So you can think of it as like a personal CRM, but it gives you context behind all of your conversations that you've had with that person. And even before you've had a conversation with them, you're able to get a report on that person. So it helps you before the call, during the call, and after the call, if that makes sense.

Anurag Kejriwal:

Visually it's way easier to explain, but you're essentially able to go into your relationship tracker, click on a person you've met with, see your different conversations you've had with them. And we're currently building a model on top of what we already use that allows it to, reason on your relationship so that in the future, let's say you've met with them 10 times, it'll understand your relationship with your past nine conversations and build your notes to that point on the 10th meeting. So it's more than just like a standalone meeting notes, if that makes sense. So I kind of to think of it as like an otter AI and like HubSpot hybrid, if that makes sense.

Demetrios:

And so are you focusing on any specific relationships? Like you said before, it was very sales focused. But is this for anybody who's getting on calls and doing anything or is it specifically for maybe customer success folks and the relationships there or product folks when they go out and they're looking at how to improve their different features or just test the waters and do some market research?

Anurag Kejriwal:

That's a great question. I've been trying to answer that myself throughout techstars and the company because I feel like it is on paper such a generalist tool and we don't want it to be a very generalist tool. We want it to be hyper focused to a specific customer segments until we grow and then we can add more later organically. We've been growing a lot towards like founders and like smaller startup teams specifically. they're having a lot of different types of meetings with a lot of different contexts and we are able to keep their relationships with one space. So founders are meeting with 20, 30, 40 different people weekly from different companies, different investors, different contexts. Recruiting has also been a great space for us where we've been seeing like recruiters in the startup teams like really using this. I tested this on a few of my own mock interviews and it was actually really interesting to see the output because with a standard note taker you get like a generic summary.

Anurag Kejriwal:

But with this we have custom templates that are 100% customizable so you're really able to get like exact domain specific information which has been really helpful even in my life. Like as a founder, I'm meeting with so many different people, completely different context. It doesn't even have to be found. Very sonnet specific. But I think it's been important to really understand how I manage those relationships.

Demetrios:

I see the recruiting one, especially if you have a potential candidate going through and meeting with five people on the startup team before they get that offer letter.

Anurag Kejriwal:

Exactly.

Demetrios:

Well, it would be nice to see what is going on with that. But I also could see the other side where I know I've been at startups and they don't let you talk when you meet someone because they don't want you to get biased by what the last person that met with them said. So that's a fascinating angle. I also think the angle of a founder a hundred percent works because you're going to be meeting with all these different Personas all the time. You look at any founders calendar and.

Anurag Kejriwal:

It'S just stuffed full with me.

Demetrios:

So it makes sense.

Anurag Kejriwal:

Exactly. And like I think even from that angle, it's been, I would say, a big challenge has been really finding that narrow target audience because, before getting into business, you want to, you seem, I want to just get everyone and target everyone so that the most people can use my product. And it's been kind of hard letting the idea go and. Because, slowly I've been trying to position the product to be more specific. So I'm but I'll lose this, segment of customers. But I've been realizing, going through the TechSource cohort and, you know, talking a lot of different customer discovery cults, I've been trying to understand, the people that would fight for keeping Sonnet are and who we want to target.

Demetrios:

Have you? That's a great question to ask or to think about who would fight for keeping it and how do you measure that?

Anurag Kejriwal:

There's been a couple ways, I would say. So our servers have been down for a couple days, like two weeks ago. And it's definitely not like a conventionally good way to think that when you get bad emails, that's a good sign. But I was actually happy that I got, a good amount of emails saying I really needed this meeting recording, but it's down. When are you able to fix that type of thing? And it was really interesting because like that. That was one of the first moments I was like our customers actually care about the product and, they're using it for an essential thing.

Anurag Kejriwal:

So that's been. I mean not to say that I want the service to be down, but it's been cool to get those emails and understand there's real people in the world that use this. So that's been one thing again, a lot of actual customer discovery, interviews. So really talking to the people who pay for our product and who don't, understanding their perspectives, why they're using it over, an alternative because, there is competition in the space. So trying to understand why did you pick Sonnet over X, Y, Z? Why haven't you paid for Sonnet? Why have you paid sort of thing.

Demetrios:

So talk to me about a few more of these features, because you said there's the templates, which I really and I haven't seen that before. What other features do you have that you've been leaning into?

Anurag Kejriwal:

Great question. So amongst our cohort as well, this has been a favorite which is like the no bot recording. So a lot of the other softwares that do this need like a meeting bot to join the call. And sometimes when you're having a one on one or something, you don't want an annoying like Audrey I bought or some kind of bot like to join. And we've been really good with that. So we have this like algorithm that allows us to not record with the meeting but and it just does two way audio through your Mac. So that's been interesting. Another thing is like with our notes themselves, we can create like visual notes.

Anurag Kejriwal:

So instead of just plain text, we have like tables, graphs, timelines, concept maps that it can create relatively hands free. You just literally just change the template to whatever you want. Which has been interesting. And I think one of the biggest like keys with our product is that like features that we didn't even envision our customers have told us about. So like with the templates, right there's infinite things you can do with it. And we have these a couple features that we think are really cool with the templates. But then customers have told us I was able to do this or like hey, can I do this? And then I experimented.

Anurag Kejriwal:

I was like you actually can. And I think all of that gets tied into the relationship tracker and that's really the key piece between our product. It's like notes are cool but relationships are better. You need to actually have a context to those notes and that's really what we've been trying to hone in while staying away from calling ourselves a CRM per se. But I think that really brings it all together.

Demetrios:

What have been some key challenges you've found when building with AI?

Anurag Kejriwal:

I think the biggest thing is audio. So working with audio and Deepgram has been a big help for this. But I think while coding anything with audio, whether it's like splicing it, streaming it, we have like different ways you can kind of hear the audio. it's a challenge for sure. And it's something I haven't done a lot before with traditional projects that I've done or even school projects that I've been on. Like audio is not really a big piece of our like learning. So really trying to understand how you can best. Because audio is sensitive.

Anurag Kejriwal:

I guess for lack of a better Word, like it's hard there's echoes that get created randomly and that's been a big challenge for us to kind of fix. And then that background noise. exactly. That's been a big thing. And then we don't have a meeting bot. So then like the way to kind of handle that two way audio stream has been difficult to kind of code up and understand as well.

Anurag Kejriwal:

Kind of I guess it's not necessarily just AI, but we currently have like a Mac app, but we don't have a Windows version yet. We have a wait list that's really growing. But the one thing, the one reason we can't make a Windows app immediately is because of that two way like audio stream. They have like a module on Swift that you can actually use, but obviously that doesn't run on Windows. So that's been kind of a struggle trying to put that in Windows. Not saying that there's not a way, but we have to kind of find a workaround for that.

Anurag Kejriwal:

In terms of like just building with AI, I think some of these like AI coding tools have been a really big help because in school and like in general, like even traditionally, we've been taught to like not kind of use this kind of stuff. It just feels like cheating in a way. Or it's like you're using stuff that's helping you without you actually doing anything. And I think there's a little bit of truth where if you're really up the deep end and like only like very dependent on AI and you can't code without it, I think that's a problem. But I've been trying to adapt and you know, use these tools in my workflow. Get like a basic MVP and then change up the code more specifically. And I think that's been a huge help in like development speed, especially with such a small team.

Demetrios:

I was going to ask about that at school. Do they help you and encourage that or do they kind of frown upon it and it sounds like in a way they're frowning upon it right now.

Anurag Kejriwal:

It depends on the teacher, to be honest. You just have to kind of be open and let them know and then kind of get their feedback. I would say generally it's not the best, like if you are using it for most classes, but some teachers are okay and accepting to that idea. So I would say it depends.