![Bryan Smith, CEO at Expedient](http://media.assettype.com/cdomagazine%2F2024-04%2F1803a38a-c0d4-4da6-9472-e8a93a73cc10%2FBryan_Smith__CEO_at_Expedient_16x9.png?w=480&auto=format%2Ccompress&fit=max)
(US and Canada) Expedient, a leading provider of cloud infrastructure and managed services, has unveiled its latest innovation: the Secure AI Gateway. This corporate-ready offering is designed to simplify access and maximize the value of generative Artificial Intelligence (AI) for enterprises and organizations of all sizes.
“The productivity boost that AI delivers is undeniable and increasingly well documented. But as more and more users embrace it in their everyday work, the risk of leaking sensitive or confidential information goes up exponentially, which represents a material risk for an organization,” says Bryan Smith, CEO at Expedient. “Secure AI Gateway gives organizations a foundation for using AI in a safe, transparent, and scalable way.”
According to a 451 Research's recent Voice of the Enterprise: AI & Machine Learning, Infrastructure study 1, security is the top concern of organizations when considering their AI/ML implementation. Additionally, the report revealed widespread concerns about reliability and ease of use.
Expedient’s new Secure AI Gateway addresses these concerns head-on. To prevent the sharing of proprietary data externally, the gateway can be configured to block the release of sensitive data or personally identifiable information to public AI models. Users securely access approved AI models through single-sign-on (SSO) authentication connected to an organization's user directory, allowing administrators to monitor usage. Role-Based Access Control (RBAC) provides organizations with the ability to control what data and AI models users have access to based on their role, and integrated observability tools deliver a comprehensive view of usage and identify security vulnerabilities.
Expedient’s Secure AI Gateway supports multiple AI models simultaneously and delivers out-of-the-box access to leading models like OpenAI’s GPT-4, Meta’s LLAMA, Anthropic’s Claude, Mistral’s Mixtral, and Cohere’s Command, along with others.
With a full suite of text and image-focused AI models available, users benefit from a single interface that they can access multiple AI resources. In addition to chat, custom applications can be built on top of the gateway, leveraging its API access, to meet the specific needs of the business.
Secure AI Gateway is our first offering in the AI CTRL family that sets the foundation for providing secure access to AI tools so that organizations are comfortable empowering their employees to unlock the benefits of AI. We have more to come throughout 2024,” Smith said. The rest of the suite will confront the hurdles companies face in adopting AI including connecting multiple data sources, improving model performance and accuracy, and hosting private AI models for when clients want to keep 100 percent of their data inside their environment. This suite starts with security and employee empowerment to eliminate “shadow AI” but also demystifies the process of integrating AI, providing clear guidance on where to begin and which technologies to utilize.
This offering assists our clients in making their first move toward responsible AI adoption,” says Bradley Reynolds, SVP for artificial intelligence at Expedient. “Our objective is to simplify AI adoption and enable clients to create applications that address their business issues with minimal coding. We envision a future that seamlessly integrates AI models and corporate data amassed over decades, helping organizations securely leverage and manage this emerging technology for maximum impact.”
Other AI CTRL offerings will deliver seamless connectivity to hundreds of data sources, sophisticated AI-ready storage for improved accuracy and performance, along with options for secure private model hosting for data that you never want to leave your infrastructure. AI CTRL will streamline data access, bolster semantic search and retrieval capabilities, and ensure robust security for sensitive data.
Key features include real-time data access from sources like databases, SaaS platforms, and spreadsheets, and the use of vector databases to enhance the efficiency of AI models. The incorporation of Retrieval-Augmented Generation (RAG) will significantly improve response accuracy and relevance while dramatically reducing hallucinations.