[Raised $8.4 million] Ep.112 - The First 100 with Suman Kanuganti, the Co-Founder of Personal.a

Suman Kanuganti is the Co-Founder of Personal.ai, an artificial intelligence that can act just like you. It is an AI version of yourself that uses personal data, facts, and opinions to create a responsive and interactive experience that others can use. Simply put, it is not a large language model (LLM) trained on all available information or a “virtual assistant or chatbot” but rather a personal language model (PLM) trained exclusively on the data that you control. Think of Personal.ai as an extension of yourself that can interact with anyone in your voice and your style. Personal.ai has raised to date $8.4 million.

Where to find Suman Kanuganti:

• Website: My AI | Create an AI Version of Yourself | Create Your Own Personal AI
• LinkedIn (12) Suman Kanuganti | LinkedIn

Where to find Hadi Radwan:

• Newsletter: Principles Friday | Hadi Radwan | Substack
• LinkedIn: Hadi Radwan | LinkedIn

If you like our podcast, please don't forget to subscribe and support us on your favorite podcast players. We also would appreciate your feedback and rating to reach more people.

We recently launched our new newsletter, Principles Friday, where I share one principle that can help you in your life or business, one thought-provoking question, and one call to action toward that principle.

Please subscribe Here.

It is Free and Short (2min).

Transcript

Let's do it. Broadcasting from around the world. You're listening to the first 100, a podcast on how founders acquired their first 100 paying customers. Here's your host, Hadi Rodwan. Good to have you on the show, Suman. How are you doing today? Nice. Nice to be here, Hadi. I'm excited to talk to you. Thank you for stopping by. I'm really excited to talk about your startup, because it has an interesting angle when it comes to the AI. But first, before we go into that, for my listeners, Suman Kanu-Ganti is the co-founder of Personnel.ai, which is an artificial intelligence that can act just like you. What this means is basically it's an AI version of yourself that uses your own personal data, facts, opinion to create a responsive and interactive experience that others can use. So we'll talk into that. We'll talk about the difference between what's an LLM, what's a PLM, which you talk a lot about in your posts. And you've raised to date around 8.4 million to bring personal.ai to life. So man, this is either extremely an interesting and exciting topic for some people or it's a scary topic for some people. So tell us, how did you come up with this? How did I come up with this? This was back in 2020 when I embarked on a new journey for the startup and the inspiration and the desire was almost like incubated over a period of like two years with the motivation of like one thing which is my innate desire to have a conversation both emotionally as well as intellectually. With a past mentor of mine of my previous company, his name is Larry Bok. You know, after he passed away because of pancreatic cancer, I had this huge posters in my office, like what would Larry do? And I always wished I had Larry's AI. So, you know, this core idea of like, you know, learning and exchanging and communication and accessing people is kind of like limited by, you know, time capacity as well as access. So yeah, so we started like embarking on this journey with... this desire to create an AI that acts and speaks not just like you, but is also grounded in your facts, in your opinions, in your memories, in your style. It's almost like true representation of who you are, but we also wanted to do it with principles in mind, which is they would own the personal data, which is the memory that they belong to. They would own the model. It's also very controlled by the user. And more importantly, the technology choices and the models that we have chosen to build is in the contrary of what the industry is very popular for today, which is the last language models. Yeah, we chose to build something totally different called personal language models. These are small models. These are grounded. These are like 120 million parameter model that are very cost effective and very scalable as well. So yeah, it took us like three years to put all the pieces together, the whole infrastructure together here to take this. you know, baby to the market and give it to people to take control of their AI life. Amazing. So if we can decipher this for our listeners, so most of the people know chat GPT. Chat GPT is essentially you look to their website, you ask them a question or set of questions and they come up with an answer based on the LLMs models that you talked about trained on data that exists out there. Today, if I go to personal.ai What's the experience like? So what's the output and how can I feed it my data so that it learns? And then where can I use personal that AI? Is it the Chrome extension? Is it the website? Is it something that follows me on my laptop? Talk to me a little bit about that. Sure, so chat GPT is a pre-trained model on the internet. So that's very similar to searching the internet or seeking the information from the internet. Personal AI is not a pre-trained model. Personal AI is a model that individual people train on themselves with their personal data that may or may not be available on the internet. So this is essentially to scale like yourself with your own knowledge, with your own memory. That's what you will experience. So when you go to personal.ai, you would experience a system where you can create your own AI. and start adding your own memories and start communicating with your own AI that essentially speaks on your behalf and speaks like you. The system is designed to have what we refer to as active memory, meaning on an ongoing basis, you can add your memory. If your AI makes mistakes, you can actually edit it and reinforce it and fix your memory as well. So the fundamental difference in here is you are training your own model that is personal to you. It's almost like your personal website versus, you know, in general, internet of data. You asked about the extensions. We do provide like full native experience and service for people to train their own AI and then share their AI in a one-on-one private chat. You can... Think about this as any other like chat GPT style kind of an experience or you know, character AI kind of an experience, but the difference is this is human to human communication with your AIs in the loop. So the more closer analogy is, it's more like WhatsApp communication or Slack communication or iMessage communication where you are communicating, you know, with people, but they have their own AI to assist them in creating replies, drafting the responses and sending it to them. And you can choose to, go in what we refer to as copilot. Copilot is something where human approves that message is accurate or edits it and sends it, which goes into reinforcement learning. Or you can go into fully autopilot mode. Autopilot would be like you can talk to somebody else, and their AI will automatically respond. And of course, there are some control settings such as personal scores and who goes on autopilot, who goes on copilot. So pretty much everything is in the hands of the person. So think of what this as like an account on WhatsApp. that account is you, plus your AI. What's the pain point you're solving with personal.ai and then what's the vision? You know, when we started off this company, we started off with a problem, right? Like I mentioned a little bit about the story. And the core problem is that people will always invent new things for creating more time for themselves, increasing access. When I say access, it's about exchange of information between people around you, like communication. And then finally, capacity. Capacity would be the information that you would remember, that you would seek, that you would process, that you would synthesize. And over the years, the way I see the world is it has been evolving based on the technology evolution for these three core things. Phone, like increasing access time and capacity, like you're learning from each other. You are emails, also communication and increasing and learning from each other. And then comes the internet. And it's not just about human to human, but also human to businesses are now, like everything is done like digitally. And now apps, you open an app, every person and every service, some form of communication, either messaging or their own company apps. So I tend to think AI is almost like the new medium, new unlock for these three issues that we desire to continuously. improve and continuously get better at, which is creating more time, increasing more access amongst people, and then our desire to just keep learning and increase our own capacity. So that's how we see it. From a practical application, an individual person, let's just say, who has graduated, who is probably working, and specifically like small business owners, let's just say authors, coaches, consultants, speakers, professionals, product managers, these people have... evolving sets of knowledge right so they would want to create their own AI to almost utilize them on a day-to-day basis to serve as an extension to their own mind create replies create drafts create draft emails create text messages for themselves so that way they can simplify price and sense and sense and set in a corporate setting in a medium-sized companies what we are seeing is there's a lot of like organizational knowledge that is packed into you know, a few executives, few VPs, marketers. And now they have an opportunity to create a second themselves, you know, for their teams to create more value or for their chief of staff to create more value. So yeah, so those are like the practical applications that are happening, but at the core, it kind of solves some of the fundamental needs that we continuously, you know, evolve as we adapt to the digital lives. Just a quick word from our sponsors and we will be back. Today I'm gonna tell you about a podcast called Beyond ERP, which has an interesting premise. It talks to C-suite executives about NetSuite implementations and cloud technology. They aim to create a valuable resource for the current and aspiring business leaders to learn and experience from them. I would urge you to check it out. It's called Beyond ERP. When you started personal.ai, how did you identify your ideal customer profile? Did you say, I want to have this product from day one for the masses? Or did you choose a niche and say, okay, let me validate the idea here. And then if that a thousand true fans adopted, then we have a winner. Yeah. It was very confusing, I will be honest. Because the idea of like having a second you is so broad that everybody would want to have a second me, right? So it's almost like if you ever wish you were a second me, then this product is for you, but there are like many people. So initially the core thesis was the capacity, the memory, like, you know, do not forget anything that makes up your own mind, like remember everything. We kind of identified and talked to like almost thousand different, you know, people who would want to be leveraging such kind of technology. And then once we had the product complete, which was very recent, like earlier this year, we took an approach of a wider net. You know, hey, look, this is what you can do. Try it out. And of course, it's early stages. And we got all kinds of different people, like productivity people, professionals, small businesses, corporations, developers. And the people that who are gaining the most value out of it are the people... who care about their personal knowledge, their personal data, their personal style. And this idea of being able to control and create the model that evolves, that grows alongside with you, which is yours. It does not belong to a company as such, it belongs to you. And those people turn out to be small business owners and medium-sized businesses right now, and individuals at these companies. Most of the small business owners, I mean specifically. authors, coaches, consultants, you know, speakers, people who have something to offer in professional life. And we're also seeing, you know, trendsetters, if you will, like, you know, people who already has like existing wisdom, they would want to like, oh, you know what? I'm just like tapped out my time and capacity and access. I want to scale myself more. Those are the people that are sticking around because once you get to a good memory, it's so magical. It's too good. But... My goal is every person should have their own personal AI. And that's the reason why there is a free version of the product as well, where you can come in, you can create, you can get your first-hand memory. And for a regular day-to-day people who probably doesn't have the scaling need, it's still awesome because you essentially have these foundational elements of who you are. What makes you as a person? What is your philosophy? What's your food preferences? What are your life choices? Experience preferences right now right now majority of this preferences and majority of this idea of who you are is tapped or locked into many different platforms who you are as a person on Twitter open table, you know Facebook Airbnb Amazon like everybody knows parts of you. Nobody knows a collection of you So yeah So at the basic people can basically have this personal AI and put all the key memories in there That makes your experience with the internet with the AI that much more personal. Makes a lot of sense. Early on, what was your acquisition strategy? What channels did you go after? What tactics, frameworks to get your first 100? We started off with podcasts. We also started off with showcasing the potential on events such as like TechCourage Disrupt, if you will, and as well leverage some of the platforms such as like Product Hunt. And one of the things that we found was we need, We were trying to identify what the PMF is. We had to go to multiple different platforms and speak to them what this platform can do so that way you get people to try it out and give you feedback. So those are some of the examples, TechCrunch Disturb, ProductHand, events, workshops, and webinars that kind of attracted different kinds of people into the system. We didn't have like huge marketing expense as much because we were still in the early stages. of the company and more interestingly it's an AI world right now right like one year ago this was not an AI world and then we have to like educate people the differences between LLMs and PLMs and you know what does it mean why do I make choices one versus other what are the differences so there is some education and awareness that naturally happen so yeah so now we are trying to be basically you know be out there and tell why for what reasons people can stick to LLMs and for what reasons people could make decision to use PLMs and the use cases are different. So I think it's okay. It's a world where a bunch of personal IIs will coexist alongside with a few LLMs. Which channel do you think was the best channel, has the highest ROI and did you do anything that is non-scalable that you only had to do it early on but you cannot do it going forward? I think major events such as Tech Runs Distrupt is like a one-time activity, one-time event. I think that's a good channel. I always encourage people to find those stages so that way you have a group of people paying attention to what you are saying. Product Hunt was definitely like a good ROI. We had good people come in. There is a community feel to it and there is an explicit understanding of what's going on with the platform. So to gather the feedback and to engage with people, to hop on the calls. I think it's a good platform. It can be ongoing, but I haven't leveraged as much yet, probably because the target market we found is not on product and maybe somewhere else. The other channels for casting like wider net, I mean, Twitter and social media still has its thing, but the problem is it cannot be just fundamentally driven by like... people or the founders inside the company, you would have to think through who is the right champion for a specific group of people and work with them to release your beta product or testing product if that's kind of where you are going. So that's the approach I would take. Amazing. What's a principle that you live by that has helped you in your journey? Principles you said? Oh, I always take pride that this is a principle-driven company, mission-driven company. One of the key things is doing right by the people, not by the platform. In other words, I think it's completely okay that we have utilized or extracted tons of value from the platforms that have been existed like over the past 20 years. But I think that now the world is different. Now there is a lot more concerns around the privacy. There is a lot more concerns around the data ownership, like how my data is being used. People want more visibility. People want more transparency. And especially in the situation of personal AI, I mean, we are talking about training a model and your memory. So we got to be aligned with doing right by the people that then comes down to how you structure your memory, how do you structure your data, how do you structure your models? And one of the core differences between LLMs and PLMs is just that. Like no memory is shared between people. Every person gets their own memory, their own index, their own model. It's a unique model that is counted from, trained from scratch. And the architecture is different. That's kind of very interesting parallel line right now. It feels like our competition is LLM, but I would consider LLM solves the specific sets of use cases, and we have personal use cases that we are going after. And there is likely a little bit of overlap of fine tune based approach from LLM to PLM. One last question, Suman. What's next for your company? Well, I want to get this baby in the hands of as many people as we can, grow the company. not just about growing in revenue, but it's also about creating and making an impact and solving the problems that people are looking for. Because people do want AI that is personal to their data, to their lives, to their knowledge. And I think these people should also not be worried or scared about AI replacing you, because at the end of the day, every human, every person has their own knowledge and their own story. So it's about leveraging and training and building their own AI for their own benefit rather than worrying about potentially a smart or large language model AI replacing their jobs. Suman, thank you for stopping by. You have an amazing story, amazing product. We'll put in the show notes links to personal.ai. Hopefully people can try it out and they would have their personal avatar there working with them. Thank you again. Where can people reach you and are you hiring? People can reach me at s.personal.ai. So that itself is our platform. That's where majority of communication now happens because now it's kind of private and I feel good about it. But you will also taste the experience of my AI. So if you want to talk to my AI, it will be the same link. And once you add friend, you can ask me for autopilot experience. Hiring is slow now, but I believe it would be talking to more people over the next month if not like early next year but if you do talented align with the mission of the company and can help out it's always better to start talking so yeah please reach out. Amazing thank you for stopping by we wish you the best of luck in your journey. Thank you so much for listening to the first 100. We hope it inspired you in your journey. If you're enjoying the podcast, please subscribe to our podcast on Apple iTunes, Stitcher, Google Play or Spotify and share it with a friend starting their entrepreneurship journey. Leave us a five star review. Your support will help spread our podcast to more viewers.