The Rise of AI Agents: From Venture Insights to the Next Frontier

Last weekend, I had the privilege to sit at a fireside chat with VC investor Tess Hau from Tess Ventures and another founder exploring potential paths for his startup. Observing the entire conversation, I experienced a founder/investor interaction from a third person’s perspective, in person, for the first time.

It was a new and eye-opening experience for me. 🤓 I couldn’t help but put myself in the shoes of that founder, brainstorming strategies alongside them. My mind was racing with ideas.

One of the highlights was Tess, a Web3 investor, discussing her investment in a Web2 AI company focused on AI agents.

We’ve covered AI agents several times in our newsletter and YouTube Channel.

It was a clear reminder of the significant role agents will play in the AI landscape.

Over the past couple of years, the world has marveled at the impressive capabilities of generative AI (Gen AI) models, such as large language models (LLMs) that can generate human-like text, images, and other content. However, the next stage of gen AI is poised to be even more transformative -

the rise of AI "agents".

AI agents represent a shift from knowledge-based gen AI tools to systems that can independently interact with the digital world, plan and execute complex workflows, and collaborate with both humans and other AI agents. These agentic systems have the potential to automate a wide range of high-complexity use cases that have historically been difficult to address in a cost- and time-efficient manner.

The key advantages of AI agents stem from their ability to handle multiplicity, use natural language as a form of instruction, and integrate with existing software tools and platforms. Unlike rule-based automation systems that can break down when faced with unexpected situations, AI agents built on foundation models have the flexibility to adapt to a variety of scenarios. And by allowing users to direct agents using everyday language, rather than requiring technical coding skills, these systems can make automation accessible to a much wider range of employees.

Source

Moreover, AI agents can leverage the capabilities of gen AI to not only analyze and generate knowledge, but also to interface with a broad digital ecosystem. An agent could, for example, tap into various software applications, search the web for information, collect and compile human feedback, and even leverage additional foundation models to complete a complex workflow.

Use Cases

To illustrate the potential of AI agents, let's explore a few use case examples across different industries and business functions:

Loan Underwriting in Financial Services: The process of assessing credit risk and preparing loan underwriting memos typically involves a highly collaborative effort, with relationship managers, credit analysts, and other stakeholders compiling and analyzing various forms of information. An AI agent system could streamline this process by breaking down the workflow into specialized subtasks, with different agents handling communications, data gathering, financial analysis, and quality assurance. This could reduce the time and effort required to complete a credit memo by 20-60%.

The credit reviewing process, with and without an AI Agent, is illustrated in the figure below.

Code Documentation and Modernization in Software Development: Modernizing legacy software systems often requires engineers to painstakingly review and document millions of lines of code, a time-consuming and complex task. AI agents could be deployed as "legacy software experts," analyzing old code, documenting business logic, and translating it to an updated codebase. Concurrently, a quality assurance agent could critique the documentation and produce test cases, helping to iteratively refine the output and ensure adherence to organizational standards.

Online Marketing Campaign Creation: Designing, launching, and running an online marketing campaign typically involves