Kindo, a platform designed for enterprise-wide artificial intelligence security management, today announced it has raised $20.6 million in new funding and acquired the open-source security project WhiteRabbitNeo.

The round was led by Drive Capital, with participation from existing investors RRE Ventures, Marlinspike Partners, Riot Ventures, Eniac Ventures, New Era Ventures and Sunset Ventures. The new funding brings the total raised by the company to $27.6 million, including a $7 million seed funding round led by Riot Ventures in September 2023.

Kindo’s platform allows enterprise businesses to access any AI model, including private, commercial and open-source, even across software-as-a-service integrations via a single interface with high visibility and security controls. The objective of the service is to do the security busywork to reduce the potential of vulnerabilities, exploits, data leaks and other issues while providing simple control and management of every AI system a company has.

“The easiest way to think about this is we’re an orchestration platform that lets enterprises of all sizes integrate any AI capabilities or any AI model that’s in the market,” founder and Chief Executive Ron Williams said in an interview with VentureBeat. “It’s in an API. You can integrate it including your own models. In IT and security, you can control who accesses those models, applications or other data sources, and what agents can talk to which models. It gives centralized control.”

Amid Kindo’s features, enterprise users gain fully centralized access to AI governance and security controls, fine-grained controls and audit logging, plus integration with more than 200 SaaS applications. All of these capabilities are supported for on-premises, cloud and hybrid environments. The platform works with many models including Google LLC’s Gemini, Meta Platform Inc.’s Llama 2, OpenAI’s GPT4 and Anthropic PDC’s Claude.

Kindo also provides complete visibility and control of AI content, from personally identifying information to regulated data. That allows companies to comply with regulations such as HIPAA and PCI-DSS by redacting sensitive data before it ever appears external to firewalls.

The company also provides proprietary and fine-tuned models based on a company’s private data that are easy to deploy and can be integrated within a company’s security perimeter. By combining enterprise data with a general model, it can become an expert on a company’s particular culture and knowledge base, greatly increasing its accuracy for knowledge work.

In addition to the investment, Kindo announced that it acquired WhiteRabbitNeo, an open-source cybersecurity AI model project. WhiteRabbitNeo is an AI model family of the same name built on the Llama 2 large language model specialized in attack and defense strategies for providing guidance and training when dealing with cyber threats.

“When I introduced the first AI model for WhiteRabbitNeo in December 2023, I mentioned that someone needed to build an AI-focused on offensive cybersecurity and that it would be best to do it in public,” founder Migel Tissera said in a statement. “This visibility provides a level of transparency and accountability to what is normally seen as a ‘dark art.’”

Tissera said the acquisition will bring much-needed financial backing to the project and deliver the technical expertise needed to integrate and expand LLMs throughout the security sector. The WhiteRabbitNeo LLM model family is planned to migrate into the Kindo platform, where it will become part of the enterprise offerings for human-in-the-loop agents and assistants. Tissera will join the Kindo team as an adviser.

“We’re supporting the open-source project and expanding the community for it,” said Williams. “It’s really good at finding security vulnerabilities in your code.”

Image: SiliconANGLE/Google Gemini