Delphi - the verified digital cloning platform 👥

CV Deep Dive

Today, we’re talking with Dara Ladjevardian, Co-Founder and CEO of Delphi.

Delphi is an AI platform that allows users to scale their expertise through personalized conversations via text, voice, and video. Inspired by Dara’s experience of wanting to preserve his grandfather’s wisdom, Delphi allows users to interact one-on-one with digital clones that replicate not just what someone knows but how they think.

Since launching, Delphi has attracted a range of users, including coaches, authors, influencers, and business leaders who want to extend their impact. The platform integrates with tools like WhatsApp, Slack, and personal websites, making it easy for users to engage with their clones in different ways.

Delphi recently secured seed funding from Founders Fund, Lux Capital, and Balaji Srinivasan, and they’re wrapping up a creator-focused round with backers like Codie Sanchez, Tiago Forte, Matthew Hussey and Nas Daily. This new funding will help Delphi further its mission to make personalized expertise widely accessible.

In this conversation, Dara talks about building Delphi, the challenges of making digital cloning intuitive, and what’s next as they push the boundaries of AI and human interaction.

Let’s dive in ⚡️

Read time: 8 mins

Our Chat with Dara 💬

Dara - welcome to Cerebral Valley! First off, give us a bit about your background and what led you to found Delphi? 

Hey! My name is Dara, and I’m the Cofounder & CEO of Delphi.

A bit about me: I was born in Houston, Texas, to immigrant parents. If you're a child of immigrants, or you know someone who is, you'll understand that you're often told to figure things out on your own, without asking for help. So, growing up, I never really had a mentor. I studied physics, computer science, and math at Georgetown, and then moved to San Francisco to work at C3 AI, where I focused on building NLP-based applications for government and oil and gas companies.

I was closely following GPT’s developments, and when GPT-3 was released, I decided to leave C3, move to LA, and start my first company, which was called Friday. Friday was an AI assistant designed to help parents with shopping—things like buying batteries, light bulbs, and other essentials that you always remember you need but often forget to actually buy. Parents could simply text a phone number, and the assistant would take care of it.

At the time, I was a solo founder. I hadn’t raised any money, and I didn’t have any advisors or mentors. Around then, I was reading a book about my grandfather, who had been a very successful businessman in Iran before the revolution. I found myself really wanting to ask for his advice—not just about his life, but what he would do in my situation, especially since I had no idea what I was doing.

Since I was already working with LLMs, I created a version 1.0 Delphi clone of my grandfather. It was incredibly therapeutic, especially because he had suffered a stroke two years earlier and couldn’t really speak. The clone turned out to be useful too—it helped me think through the kind of business I wanted to build.

Eventually, Friday was acquired, and I moved to Miami to work at OpenStore. While I was there, I became fascinated with how we learn from others. We read books, watch YouTube videos, listen to podcasts, follow people on Twitter or LinkedIn—but none of that information is personalized to us. It doesn’t adapt to our changing life circumstances, and it doesn’t hold us accountable. Studies in education show that one-on-one learning is the most effective, but most people can’t afford a $10,000 coaching session, nor are they lucky enough to have access to top-tier experts.

On the flip side, there are experts who want to scale their knowledge. They write content or hire teams to train others in their way of thinking, but that often dilutes their unique essence. I met my co-founder at OpenStore, and together, we decided to start Delphi—a digital cloning platform that captures your expertise and allows you to scale one-on-one, personalized conversations over text, voice, and video.

Give us a top level overview of Delphi - how would you describe the startup to those who are maybe less familiar with you? 

The best way to put it is that we're scaling one-on-one interactions. This is really the next medium of communication and education. When you read a book, it's passive. Maybe you’re imagining a conversation with the author as you read through different passages, but now with Delphi, it actually responds. It asks follow-up questions. It's a whole new way of interacting with someone else's knowledge.

Who are your users today? Who’s finding the most value in what you’re building with Delphi? 

So, there are definitely phases to Delphi in terms of who we're focusing on because there are two groups of users—supply and demand. In the long term, this is a marketplace. You have people who want to learn from or talk to someone, and then you have those who want to do the talking and teaching. We started by focusing on the supply side. I’d say about 70% of our customers on the supply side are what I call knowledge-based professionals. These are coaches, authors, influencers, course creators, experts, and thought leaders. The other 30% are business owners—CEOs who are looking to scale themselves for customers or new employees who want to learn how the business works or how they would handle specific situations. We also have a small portion of users who are using Delphi for legacy, family preservation, and digital immortality. I think that’s going to continue to grow over time, but we purposely didn’t start there because it’s tough to move out of that market once you’re in it.

There’s been an explosion of interest in multimodal and agentic workflows. How has that shaped the way you’re thinking about building Delphi? 

We support all mediums of communication on Delphi - text chats, voice calls, and now video calls…who knows, maybe holograms will be next. Obviously, we have a lot of product challenges—like making it super intuitive to create a clone and ensuring it’s genuinely useful—but the main technical challenge is capturing not just what someone knows, but how they think. Can we accurately predict human responses in entirely new situations and make the clone behave as the person would? 

The concept of an AI clone is definitely pretty unique, even by today’s standards. How have you earliest users engaged with the idea - any surprises? 

Brand is super important for Delphi, kind of like how it is for CAA, the agency in LA. When you walk into the CAA office, the ground is granite, it sparkles, and you immediately feel like you’re in a special place. You think, “I want CAA to represent me.” Delphi has a similar vibe because people are coming to us with their data and saying, “Delphi is going to represent my identity to my audience.” So trust is crucial. At Delphi, only the person themselves can create a clone. I manually verify every single person that comes through, and any data used to train a clone is owned by them—it isn’t shared with any external providers. That trust is so important to us.

It’s been really cool to see people interacting with these clones and getting real value out of it. Some people will have a 30-minute call with a dating coach clone, and then they’ll come back later. It’s interesting when someone talks to my clone and then emails me, saying, “Hey, I spoke to your clone, would love to hop on a call.” When we finally get on a call, our conversation is almost deeper because they already know so much about me. 

There’s definitely some pushback against cloning, and I think part of it is the term we use. Some people are a little afraid or even offended, like, “Why would I talk to a clone? That’s almost offensive.” But the way I see it, a clone is just an artifact of a human being, the same way a book is. You’re just experiencing a new medium of their mind. It’s not meant to replace them; if anything, it’s providing access where there wasn’t any before, and it’s enhancing that relationship.

Funnily enough, some customers are actually raising money to build entire 1-1 businesses, just on their clone.

How do you view the split between consumer and business interest in a platform like Delphi? 

The demand side use cases are really driven by the supply side. For example, a dating coach can create their clone, monetize it, and it becomes a valuable tool for consumers. We’ve started by focusing on the supply side because we need to prove that cloning can actually be useful for people before we create a consumer app. To build a consumer app, you'd typically have to do what Character AI did, which is creating bots of people without their permission—and I don’t think that’s the right approach for this business. Now, we're starting to double down on our efforts to develop the consumer platform, where you have a goal and can be recommended the right people to talk to for your specific objectives. So yeah, the supply and demand sides are very much intertwined.

How do you plan on Delphi progressing over the next 6-12 months? Anything specific on your roadmap that new or existing customers should be excited for? 

The number one thing we're focusing on is clone proactivity. This means having clones that can intelligently follow up with you after a conversation and keep you accountable. Right now, clones rely too much on consumer intent, which doesn’t make them much better than a course or a book. But people want guidance toward their goals. So, instead of just reading a book on how to be more present, you could subscribe to that author's clone journey, where the clone keeps you accountable, messages you daily, hops on a call, and makes sure you’re applying the knowledge. I'm really excited about this because it’s going to have a real impact on consumers and open up new monetization channels for clone creators.

We're also revamping the consumer platform so that as you talk to clones, each one becomes more personalized to you. You'll be able to get recommended topics to learn from based on your goals and clones to learn from based on what you're trying to achieve. Those are the two main focuses right now. Beyond that, we're really zeroed in on improving the product. The difference between what we have now and the goal of actually capturing how someone thinks is almost infinite. We're essentially trying to recreate a human, so it's going to take a while to get to that point where you think, "Holy shit, this really is the person."

How do you measure the impact that Delphi is having on your users? Do you tend to look at conversation length, depth of interaction etc.? 

It really varies depending on the customer. Some customers prefer shorter conversations where they can get straight to the answer, while others, like those in coaching or dating, might want longer, more in-depth interactions. We usually know pretty quickly if a customer is unhappy with their clone—if they keep paying and continue using the service, that’s a good sign their clone is representing them well. If it wasn’t, they’d probably stop using us. 

From a technical perspective, we're working on quantitative evaluation metrics to track how accurate and effective the clones are. But honestly, we base a lot of it on customer satisfaction. When we first started, we struggled to get people on board, and when they did create a clone, many would say, “This sucks, this doesn’t sound like me,” and then leave. But we don’t see that happening as much anymore, and we hope to keep it that way.

Lastly, tell us a little bit about the team and culture at Delphi. How big is the company now, and what do you look for in prospective team members that are joining?

We're a small team of nine people—it's me, my co-founder, four engineers, a designer, our head of Ops, and an intern. Everyone here is super scrappy and motivated, with a bit of a chip on their shoulder. We're all based in San Francisco, working mostly in person, and while six or seven-day workweeks aren't required, you'll find that a lot of us are doing just that. The mission drives us, and the in-person culture really fuels our ambition to be the best. We've also got a strong focus on learning—there's a book club for both the engineering and marketing sides, which I think is pretty unique to what we've built here.

As for hiring, there's one role that really stands out right now, though we're always open to the right value-aligned people. The main role we're looking for is what I call an Applied AI Engineer. This role is all about helping us improve the "clone brain." It’s a high-ownership position—not for someone looking for mentorship, a nine-to-five, or foundational research and model training. It's perfect for someone who’s excited about applying AI to capture someone's mind and enable personalized expertise. This role will need a mix of creativity, knowledge of knowledge graphs, LLMs, retrieval architectures, and things like that. 

Conclusion

To stay up to date on the latest with Delphi, learn more about them here.  

Read our past few Deep Dives below:

• Zilliz Cloud is powering vector search for the enterprise 🔋

• Kolena aims to redefine quality standards in ML 🎛

• Pinecone Serverless is turbo-charging your time-to-prod ⚡️

• Cleanlab is automating data curation at scale 🌐

If you would like us to ‘Deep Dive’ a founder, team or product launch, please reply to this email ([email protected]) or DM us on Twitter or LinkedIn.

Join Slack | All Events | Jobs