transcribe

How Factorial Built an Agent Across 25+ Products

Mastra · 2m · transcribed 2h ago
More from Mastra Business
𝕏 Share ▶ YouTube 📥 PDF 🤖 .md

Section Insights

# 0:00

Introduction to Factorial and AI Integration

What is Factorial's mission and how does AI fit into it?

Factorial aims to help small and medium businesses automate mundane tasks, allowing them to focus on their customers. With the advent of AI, Factorial seeks to integrate AI into its products to enhance functionality and streamline workflows.

  • Factorial was founded to automate tasks for small and medium businesses.
  • The integration of AI is central to Factorial's strategy for product development.
  • AI will help unify the various products offered by Factorial.
# 0:32

Challenges in AI Implementation

What challenges does Factorial face in implementing AI?

Factorial faces challenges related to context management and ensuring that AI operates with the same permissions as the user. The goal is to create an AI experience that feels like working with a coworker.

  • AI must respect user permissions to function effectively.
  • Context management is a critical challenge in AI development.
  • The aim is to create a user-friendly AI experience.
# 1:04

Choosing the Right Framework for Development

How did Factorial choose its development framework?

Factorial found a balance in the Mastra framework, which offers meaningful building blocks without being overly complex. This allowed for rapid development while maintaining a clear understanding of the underlying processes.

  • Mastra provides a balanced framework for product development.
  • The framework allows for rapid development while ensuring clarity.
  • Factorial is focused on building a robust product without unnecessary complexity.
# 1:36

Evolving Product Development with Mastra

How has the relationship with the Mastra framework evolved?

As Factorial's product matured, the team increasingly adopted Mastra's native solutions over their own homegrown methods, finding it easier and more efficient to align with the framework's capabilities.

  • Factorial is transitioning to use more native solutions from Mastra.
  • Adopting Mastra's approaches has streamlined development.
  • The team is aligned with Mastra's updates and innovations.
# 2:08

Vision for the Future of Factorial's AI

What is Factorial's vision for its AI product?

Factorial aims to develop its AI product, referred to as 'One', into a comprehensive enterprise AI solution that businesses can trust and rely on, akin to a 'Claw-like' system tailored for companies.

  • Factorial's AI product is currently in its early stages with significant daily usage.
  • The vision is to create a trusted enterprise AI solution.
  • Factorial seeks to enhance its AI capabilities continuously.

Transcript

0:00 I'm Ilya Zayats, I'm CTO at Factorial. Factorial was born 10 years ago with a really simple idea in mind. We wanted to help small and medium businesses automate boring things so our customers could focus on what really matters: their customers. And when we saw that, okay, AI is here to stay, it's not a toy anymore, we thought, can we create something? Can we put AI in the center of all the building blocks that we created to cover possibly all the workloads?

0:24 AI in the center for us means that the magic of Factorial being one product for the customer is there. We have more than 25 products, so AI needs to be able to understand from one question that maybe you need to connect the data across six products, seven products. The next big thing for us is permissions. It's obvious that AI needs to use the same permissions as the actor who is using this AI. The hard part for us is the context management.

0:51 We wanted to create something that could feel like your coworker. We wanted to build a harness for our product. And I started looking at all the frameworks around, and I found two spectrums. Either something extremely high level that kind of puts you in the church of their thinking and how you're supposed to build, or something extremely low level where you need to wire everything together. And Mastra was, at that moment, in the middle of that. So it was a perfect balance.

1:18 It has all the building blocks. They are meaningful enough for you to build a proper product, but they are also not so high level that you don't understand what's happening inside. We started developing on top of it really fast, and at some moments we found that, okay, looks like we need to flex the abstractions too much, or reinvent something that was part of the framework. But the more time passed, the more our product matured and the more Mastra matured.

1:43 I was seeing us rip out our own homegrown solutions and switch to the more native Mastra approaches, because in the end that was just easier than continuing to reinvent our own wheel. What I'm also seeing right now is that every time we're thinking about a new feature, like, hey, it would be cool to add durable agents now, and then, just two hours later, I see on Twitter, whoa, there's a new Mastra release with durable agents. Okay, looks like we're thinking in the same direction.

2:09 One right now is open to all our 17,000 customers and it has thousands of iterations daily, covering almost the whole product surface that we have. But this is kind of the first step. For now, we're seeing, okay, we have a harness on the product, how can we make it even more? So our vision for One is we want to turn it into, I know it's a little bit hype, a Claw-like one, so in the end it's a Claw but for your company.

2:35 So it's really an enterprise AI for the company, one you can totally trust and that can work for you.

Summary

Ilya Zayats, CTO of Factorial, discusses the company's evolution over the past decade, focusing on integrating AI to enhance their suite of products aimed at automating tasks for small and medium businesses. He emphasizes the importance of context management and permissions in AI, as well as the adoption of the Mastra framework to streamline development and improve product functionality.

- Factorial aims to automate mundane tasks for small and medium businesses, allowing them to focus on customer engagement.
- The integration of AI is central to Factorial's strategy, enabling seamless interaction across over 25 products.
- Context management and user permissions are critical challenges in implementing AI solutions.
- The Mastra framework was chosen for its balanced approach, facilitating rapid development without sacrificing understanding of underlying processes.
- Factorial has transitioned from homegrown solutions to more native Mastra approaches as their product matured.
- The company’s AI product, "One," is currently available to 17,000 customers and undergoes thousands of iterations daily.
- Future goals include evolving "One" into a trusted enterprise AI solution, akin to a Claw-like system for businesses.

Questions Answered

What is Factorial's mission and how does AI fit into it?

Factorial aims to help small and medium businesses automate mundane tasks, allowing them to focus on their customers. With the advent of AI, Factorial seeks to integrate AI into its products to enhance functionality and streamline workflows.

What challenges does Factorial face in implementing AI?

Factorial faces challenges related to context management and ensuring that AI operates with the same permissions as the user. The goal is to create an AI experience that feels like working with a coworker.

How did Factorial choose its development framework?

Factorial found a balance in the Mastra framework, which offers meaningful building blocks without being overly complex. This allowed for rapid development while maintaining a clear understanding of the underlying processes.

How has the relationship with the Mastra framework evolved?

As Factorial's product matured, the team increasingly adopted Mastra's native solutions over their own homegrown methods, finding it easier and more efficient to align with the framework's capabilities.

What is Factorial's vision for its AI product?

Factorial aims to develop its AI product, referred to as 'One', into a comprehensive enterprise AI solution that businesses can trust and rely on, akin to a 'Claw-like' system tailored for companies.

© transcribe · For agents Built with care and craft by Gokul Rajaram