Brevian Inc., a startup that provides artificial intelligence agents for tasks such as investigating data breaches, today launched with $9 million in funding.

TechCrunch reported that the round was led by Felicis Ventures. The deal is said to mark the venture capital firm’s largest-ever seed investment.

Large language models are customizable, which allows users to adapt them for tasks that may not be supported well out of the box. A company could, for example, equip an LLM with the ability to field help desk tickets by giving it access to a technical support knowledge base. Models customized in this manner are often referred to as agents.

Sunnyvale, California-based Brevian offers a platform that allows workers to create AI agents without writing any code. The platform currently focuses mainly on automating manual tasks in two areas: cybersecurity and customer support.

Determining whether unusual network activity may be the result of a breach often requires administrators to analyze logs from multiple systems. Manually retrieving and analyzing those logs can take a significant amount of time, which slows investigations. Brevian says its platform makes it possible to create AI agents that automatically retrieve data about potential breaches. 

Cybersecurity teams can interact with the agents through a ChatGPT-like interface. An administrator could, for example, ask a Brevian-powered AI to retrieve information about recent connections to external networks. It’s also possible to filter the data: The agent could be instructed to only display connections from the last 24 hours that transferred over 500 megabytes of data.

Customer support teams must likewise review large amounts of information as part of their day-to-day work. Troubleshooting a server malfunction, for example, may require a help desk representative to check multiple technical guides as well as logs from the affected data center. Brevian makes it possible to create AI agents that can automatically retrieve customer support information.

According to the company, its platform is equipped with a “reasoning engine” that allows it to not only fetch data but also run complex analyses. A built-in moderation mechanism prevents the underlying AI algorithms from generating unsafe responses. Additionally, responses are created in a way that makes it relatively simple for users to check their accuracy.

Brevian has also equipped its platform with cybersecurity controls. According to the company, administrators can regulate which worker may access a given AI agent and how. For added measure, Brevian includes an audit tool that logs user interactions with agents.

The company will reportedly use the proceeds from its new seed round to hire more employees. In conjunction, Brevian plans to accelerate its product development efforts. The company intends to expand its platform’s focus beyond cybersecurity and customer support to other use cases. 

Image: Freepik