Announcing our investment in Lafayette, IN based Prediction Guard, the AI system that you can deploy privately in your company’s own infrastructure.
As we approach the three-year anniversary of OpenAI publicly launching ChatGPT-3.5 (November 30th, 2022) and accelerating the AI supercycle, companies from early-stage startups to Fortune 100 enterprises are still grappling with a variety of concerns that hinder them from implementing AI into their internal operations and products.
Hesitations arise from concerns that using large language models (LLMs), particularly closed-source ones like those from OpenAI or Anthropic, could jeopardize proprietary data and customer information. Companies worry these model providers might deliberately or accidentally train on, expose or confuse the ownership of confidential, sensitive information. Further, there is a risk that an AI solution, such as a customer support chatbot, could: 1) cause harm to the company’s brand or reputation by providing inaccurate or problematic outputs due to hallucinations and toxicity; or 2) open up the company to risk from malicious actors that could trick the AI into disclosing intellectual property (IP) or personal identification information (PII).
For all these reasons, legal, cybersecurity, finance and compliance departments often slow down or block the process of AI adoption. This response is due to the inherent risks and concerns associated with AI, which end up producing a tension between engineering departments (who are eager to start building magical AI features) and compliance/ GRC (who want to maintain their data security policies).
Prediction Guard solves these pain points through its private and secure AI platform that allows companies to safely deploy AI behind their own firewall. This enables enterprises to leverage AI technology without exposing sensitive data or intellectual property to third-party models. Designed for regulated industries like healthcare, manufacturing, transportation and government, Prediction Guard empowers organizations to innovate while meeting strict compliance requirements.
The company is led by Daniel Whitenack, an experienced entrepreneur known for his tenure as co-host of the Practical AI podcast. Under Daniel’s leadership, Prediction Guard has secured strategic reseller partnerships with industry giants including Intel. Through these partnerships, Prediction Guard has developed their platform for real enterprise scenarios where companies might not have expensive GPU clusters. Prediction Guard is able to be deployed to NVIDIA GPU, CPU and GPU-alternative hardware in any on-prem, hybrid or cloud VPC environment.
“When I first met Daniel in 2017, I was super impressed with both his deep technical background but also his ability to communicate clearly and build trust. However, I couldn’t have predicted where this would lead,” said Victor Gutwein, Managing Partner. “Over the years, Daniel’s robust, multisided understanding of AI problems for enterprise led him to create one of the most critical ‘picks-and-shovels’ solution for this rapidly growing technology to thrive alongside corporate use cases.”
Their platform allows companies to deploy AI that aligns with NIST and OWASP best practices for AI security, and it offers all the core AI building blocks: LLMs, vision models, embeddings, document extraction, transcription, safeguards and more. Prediction Guard ensures that these models are securely managed and monitored, making enterprise-scale AI both accessible and safe.
We believe Prediction Guard is uniquely positioned to solve one of the most pressing issues in AI adoption—trust. Their private, on-prem AI solutions enable companies to leverage the latest advancements in AI without compromising security or compliance. As AI continues to transform industries, the demand for secure and private deployments will only increase, and Prediction Guard is leading that charge. Their founder-market fit, strategic partnerships and clear understanding of the enterprise landscape give us strong conviction that Prediction Guard will be a dominant player in secure AI for regulated markets. We're thrilled to partner with Daniel and the team as they unlock the potential of AI for the most sensitive use cases.
“It’s a privilege to partner with M25, who has an incredible reputation nationwide and is working to promote innovation that is happening here in the Midwest. We believe that our geographic location is strategic to serve industries with a strong presence in the region—like advanced manufacturing, healthcare, life sciences and logistics—with AI solutions that fit their security and compute constraints. M25 has been a key partner in helping us fulfill this vision,” said Daniel Whitenack, CEO & Co-Founder at Prediction Guard.
About Prediction Guard: Prediction Guard is a best-in-class AI system that you can deploy privately in your company’s own infrastructure. The system aligns with AI security best practices (from NIST, OWASP, etc.) out-of-the-box and includes all of the AI building blocks that your developers need to build transformative AI functionality (LLMs, vision models, embeddings, document extraction, safeguards, etc.). With Prediction Guard, you have the configurability to privately deploy the specific AI models that make sense for your use cases and industry (without passing any data through to third parties). You can then monitor and govern those AI models in a manner that fits your regulatory-burden and security posture. For more information, please visit www.predictionguard.com.
Learn more: Lafayette-based Prediction Guard secures $3.7M in seed funding for AI platform
About M25: M25 is an early-stage software-focused venture firm based in Chicago, investing solely in tech startups headquartered in the Midwest. Since launching in 2015, M25 has become the most active investor in the region, quickly becoming the preferred seed investor for the next generation of Midwest unicorns. Portfolio companies include Kin Insurance, Loop Returns, Astronomer, Branch Pay, Authenticx and more. For more information, please visit www.m25vc.com.
