SafeBase uses AI to Automate Software Security Reviews for Faster and More Efficient Detection

Entrepreneurs Al Yang and Adar Arnon, who met at Harvard Business School, share a common interest in cybersecurity. As they observed the evolving business climate, they recognized the growing need for improved security processes. According to Arnon, "Security’s importance has increased exponentially… it’s non-negotiable for technology buyers."

This realization sparked the idea of SafeBase, which was accepted into Y Combinator’s accelerator program during the pandemic. The company helps customers fill out security questionnaires, typically a governance and compliance requirement before buying new software.

Table of Contents

The Problem with Security Questionnaires

Security questionnaires can be a tedious process, taking teams weeks to months to complete for more complex pieces of software. Arnon argues that SafeBase can save time through automation – and AI. The company employs AI models "specifically trained on security documentation use cases" to read, interpret security information and questions, and automatically respond to security questionnaires.

"Our platform takes the pain out of the cumbersome security review process by empowering security, governance, risk, and compliance teams," said Arnon. However, a skeptic might wonder about the accuracy of these AI models. According to Arnon, they are superior due to a "mix of large and small language models" that deliver "greater answer coverage."

Beyond Automation: Custom Models and Rules-Based Behavior

In addition to custom models, SafeBase provides an engine that allows companies to assign "rules-based behavior" for customer access. The platform also offers dashboards that show insights and analytics on the company’s security posture.

While SafeBase is not the only vendor offering tools to automate security questionnaires and reviews, Arnon seems unconcerned about competition from rivals like Conveyor, Kintent, and Quilt. He points to SafeBase’s 700-company-strong customer roster, which includes Palantir, LinkedIn, Asana, and Instacart.

Massive Growth and a Strong Customer Roster

SafeBase has seen massive growth in the past couple of years, with customers "loving the product" and adoption continuing to accelerate. Arnon attributes this success to increased visibility across its vendor network as more high-volume customers launch trust centers that replace the need for tens of thousands of manual security reviews.

The company is based in San Francisco and has 55 employees. The Series B round was led by Touring Capital, with participation from strategic investor Zoom Ventures (Zoom’s corporate venture arm), NEA, Y Combinator, Comcast Ventures, and Cerca Partners, as well as angels including former Salesforce chief trust officer Jim Alkove.

Funding and Expansion

This latest funding brings SafeBase’s total raised to over $50 million. Arnon plans to allocate a significant portion of the funds toward expanding the team.

"SafeBase saw massive growth in the past couple of years," said Arnon. "Customers love the product, and adoption continues to accelerate. The company benefits from increased visibility across its vendor network as more and more high-volume customers launch trust centers that replace the need for tens of thousands of manual security reviews."

Conclusion

SafeBase’s success in automating security questionnaires is a testament to the growing importance of cybersecurity in today’s business climate. With AI models specifically trained on security documentation use cases, SafeBase aims to save time and reduce the complexity of security reviews. As the company continues to grow and expand its team, it will be interesting to see how they address the challenges of ensuring accuracy in their AI models.

About the Authors

Kyle Wiggers is a senior reporter at TechCrunch with a special interest in artificial intelligence. His writing has appeared in VentureBeat and Digital Trends, as well as various gadget blogs.