AI Agents Examples: Relevance AI Co-founder Tells All
29 Nov, 2024
In this episode of the Curiosity Code podcast, Alex and co-host Ilya delve into the world of AI agents with special guest Jacky Koh, co-founder and co-CEO of Relevance AI. Jacky introduces the concept of AI agents, explaining how they transcend traditional AI automation by performing complex, multi-step tasks and enabling collaboration between specialized agents. The discussion highlights the transformative potential of AI agents in automating repetitive tasks across industries, particularly in sales, marketing, and customer support, while freeing up teams to focus on strategic decision-making. Jacky emphasizes the importance of integrating traditional and dynamic AI solutions, building robust multi-agent systems, and leveraging advanced knowledge retrieval systems (RAG) for accurate decision-making. The episode also explores Relevance AI's innovations, including a tool-building platform, multi-agent collaboration, and advancements in data privacy and security, essential for enterprise clients in industries like fintech. The conversation concludes with insights into the future of AI agents, their applications in business automation, and their potential to revolutionize workflows. This episode is packed with expert perspectives and practical advice for anyone curious about the cutting edge of AI and its impact on business operations.

Alex: Hi everybody and welcome to the podcast, Curiosity Code podcast. That was weird. Okay. Let's redo it again.

Ilya: Words, take it easy, take it easy.

Alex: Hi, everybody, and welcome to another episode of the Curiosity Code Podcast. Today is going to be quite a special episode. First of all, I'm co-hosting this episode with our partner and my close friend, Ilya. Hi, Ilya. And second of all, we'll have a special guest, and I'll explain why we will have a special guest in a minute.

Ilya: Hey everyone.

Alex: Today, everybody's talking about AI, and AI is such a generic term. We can divide it into different sub-aspects of AI, like LLMs and agents and automations and stuff. But what we hear more and more often these days is everybody talking about AI agents. So today we're going to explore that subject. We'll talk about the differences and what the AI agent is. And yeah, let's dive in. Ilya, you're doing a PhD in that subject, and you've been playing with this technology for a long time from different aspects—from marketing, technology, and business. What are your feelings about this technology? What is that?

Ilya: Sure, sure. Thanks, Alex. I'm actually working on my PhD thesis. I wouldn't say it's connected to AI in a broad sense. It's rather about processing environmental data. I'm pretty much curious about satellite data, earth observation, and how we can get insights from the data that we get from space and predict some things that are happening on our planet. But that's perhaps a topic for the next podcast episode. Today, we are rather focusing on AI agents. Actually, people are just getting to know what AI agents are. There is a shift happening these days from AI automation to AI agents. I think we are going to talk a lot about the differences between AI automation, AI agents, and some generic ChatGPT questioning that you do via the user interface. When I first started to play with AI agents, I was mind-blown. It shifted my perspective completely because of what they can do compared to previous tools.

Alex: Yeah, let's narrow it down to a use case. Let's explore use cases you've seen that showcase the success of AI agent applications in FinTech. Let's explore that.

Ilya: Absolutely. We've got a customer for whom we built a platform for market data analysis. I think you can better explain all the nuances since you closed that customer and have been working with them for a long time. You know all the details. But there are many cases where AI agents shine, especially in sales and marketing automation. AI agents are enabling smarter and deeper insights rather than shallow outreach automation. People are moving past fake personalization, like spamming LinkedIn inboxes, and using AI agents to create meaningful connections by analyzing vast amounts of data, such as social media profiles and company information.

Alex: Now, what's important in this conversation—and I think we will try to discover more in this direction as the episode progresses—is the difference between AI automation and AI agents. Can you explain the distinction between the two from your perspective?

Ilya: Absolutely. It’s a paradigm shift. AI automation typically follows predefined logic and processes, while AI agents operate more dynamically. Agents can handle complex decision-making by building thought chains and collaborating with other agents to solve problems. There’s reflection, planning, and generation involved. This flexibility is what sets them apart from simple automation systems.

Alex: Yeah, let’s explore that in a bit because we have our special guest waiting in the lobby room. Today, I’m happy to have Jacky Koh, co-founder of Relevance AI, as our special guest. Let’s bring him in now.

Jacky Koh: Hi, Alex, Ilya. Nice to meet you. Thanks for having me on the show.

Alex: Nice to meet you too. Can you start with a short introduction about yourself and your company?

Jacky Koh: Sure. My name is Jacky Koh, and I’m the co-founder and co-CEO of Relevance AI. I come from a data science and machine learning background, having led various ML teams and built mobile apps with my co-founder that reached millions of users. At Relevance AI, we focus on enabling AI agents to collaborate and tackle complex tasks for industries like sales, marketing, and customer support. We’re based in Sydney and San Francisco, with a team of about 40 people.

Alex: AI seems to be everywhere these days, almost like a buzzword. How do you see businesses moving from traditional automation to AI agents, and how is Relevance AI bridging the gap?

Jacky Koh: It’s not about replacing traditional automation but combining it with dynamic AI. AI agents excel at knowing when to use traditional automation for specific tasks and blending both seamlessly. For example, agents can decide when to query Google for information, automate parts of a workflow, and collaborate with other agents for more complex tasks. The key is to integrate and not solely rely on either.

Alex: That’s a great perspective. Ilya, over to you. I know you’ve been working closely with Relevance AI. Do you have questions for Jacky?

Ilya: Absolutely. First of all, Jacky, I’ve been amazed by how quickly we could create AI agents on your platform. Within hours, we had an MVP ready for a client, and they were blown away. Can you share insights into how Relevance AI continues to evolve and support customers?

Jacky Koh: Thanks for the feedback! We’re constantly refining the platform to make it powerful yet user-friendly. Features like multi-agent systems and advanced knowledge retrieval (RAG) are areas we’re enhancing to help customers build collaborative agents that deliver real value. We’re also focused on making tools and APIs more accessible so agents can perform diverse tasks efficiently.

Alex: A quick question about multi-agent systems: Do you think these systems could replace entire departments like sales or marketing in the future?

Jacky Koh: Not replace but augment. Multi-agent systems handle repetitive tasks, allowing teams to focus on strategy and creativity. For example, marketing agents can tag data automatically, and sales agents can streamline outreach. The human element of strategic thinking remains irreplaceable.

Alex: Makes sense. What about data privacy? How does Relevance AI ensure data security, especially for enterprise clients?

Jacky Koh: Data privacy is a top priority. We use secure models like Azure OpenAI, and customer data is not used for training. We also comply with standards like SOC 2 and GDPR, and we’re continuously improving security measures to meet enterprise requirements.

Alex: That’s reassuring. Thanks, Jacky, for an insightful conversation. It’s been a pleasure having you on the show.

Jacky Koh: Thank you for having me. It’s been great speaking with you both. Take care!

Alex: Thanks, Jacky. And to our listeners, don’t forget to subscribe and reach out if you have any questions about AI agents or Relevance AI. See you in the next episode. Bye!