Located in Sunnyvale Brevian aims to simplify the process for business users to create personalized AI agents. At present, the company is concentrating on support teams and security analysts, which have clearly defined use cases and training sets. However, there are plans to broaden into other areas in the future. Brevian is emerging from stealth mode today and revealing a $9 million seed funding round, as shared exclusively with Eltrys.
Vinay Wagh, the CEO, and Ram Swaminathan, the CTO, founded Brevian, also known as BREV/N. Both of them brought unique backgrounds to this startup venture. Wagh held the positions of product director at Databricks and head of products at Bracket Computing in the past. While at Databricks, he focused on preparing the company for acquisition by major corporations and introducing its serverless product, Databricks SQL Serverless. “That was released in February, and I was observing all the developments,” Wagh mentioned. Upon encountering ChatGPT and similar technologies, my immediate thought was: how can this be utilized in a business setting? Immediately, it became clear that all these missing pieces would be necessary. There is great potential, but without solving this particular issue, the outcome may not be successful. That was the motivation to go out and explore—there are things happening. After that, I was introduced to Ram [Swaminathan] by a mutual acquaintance.
Wagh focused primarily on the data aspect, whereas Swaminathan’s expertise lies in computer science and machine learning. Having a background in academia, he dedicated nine years to Bell Labs and another 17 years to HP Labs, focusing on a wide range of projects, including coding theory, cryptography, and the early stages of machine learning. After that, he moved to LinkedIn, leading the AI trust team until he departed in late 2022 to become Brevian’s CTO alongside Wagh.
“I lacked a product sense,” he shared with me. “You write papers, you write patents—but can you truly develop a product that can make a significant impact on a wide audience?” LinkedIn was a prime target: by participating, you contribute to safeguarding a platform that provides opportunities and connections for individuals. That’s all. This has a significant influence on society. I thoroughly appreciate collaborating with product teams and engineers, orchestrating all these elements to truly generate value.
Security is considered a fundamental aspect of all enterprise projects involving generative AI, according to the team. It is crucial for AI vendors to prevent any leakage of personal information. Upon the company’s launch, numerous enterprises opted to prohibit the use of ChatGPT within their organizations due to concerns about safeguarding sensitive data from being shared through chat interactions.
During its initial stages, Brevian primarily concentrated on security. “We rapidly developed models for detecting personally identifiable information,” Wagh explained. “Observing the rise in prompt injection attacks, we developed a system focused on detecting prompt injection based on intent.” What we did at that time was gather all the prompts that disrupt alignment, and then we tested a model on all of them, successfully identifying each one.
It is interesting to note that even now, businesses are concerned about security when it comes to large language models, but they are unsure about the specific nature of their concerns beyond potential leaks. The team soon recognized that the main obstacle hindering enterprise adoption was not primarily security but rather the difficulty of developing systems that effectively addressed their actual needs.
Our goal was to empower business users in the enterprise to leverage AI to streamline their daily activities. All of that aligned with the vision to broaden beyond security and develop these AI agents.
Brevian’s seed round was spearheaded by Felicis partner Jake Storm, setting a new record for Felicis’ largest seed round investment. A year ago, there was a lot of buzz about AI infrastructure, but not many were discussing applications or the crucial aspect of securing them. 2023 was the year of AI infrastructure. 2024 marks the era of AI applications, and it seemed like they were significantly advanced,” he shared with me.
Through this funding round, Brevian is focused on speeding up product development and enlarging its team to keep up with the increasing customer demand via its early release program.