H2O AI and Dell Technologies
H2O is validated and pre-integrated on the Dell AI Factory with NVIDIA, giving Dell customers generative and predictive AI that runs on-premises or air-gapped. Organizations can fine-tune or ground domain-specific models on their private data with multimodal RAG and agentic workflows, from exploration through inference. The solution is available in the Dell Customer Solution Center.
Partnership overview
What Dell customers get
On-premises and air-gapped, purpose-built AI applications for Dell customers: H2O runs on Dell AI Factory with NVIDIA, validated and pre-integrated for scale and deployment.
- Agentic AI: use private, protected data to fine-tune or apply retrieval-augmented generation (RAG) for domain-specific LLMs in multi-agent systems.
- Reduce TCO: choose on-premises or hybrid deployment of applications on Dell AI Factory.
- Accelerate time to production: pre-integrated technologies.
- End-to-end AI solution: use-case development from exploration to inferencing.
“Enterprises are demanding AI solutions that are trustworthy, cost effective and purpose-built for domain specific use cases, and with H2O.ai, we're delivering on that demand. The Dell AI Factory with NVIDIA and H2O.ai helps organizations deploy AI on-premises with unmatched control and cost efficiency.”
Chad Dunn
Vice President, Product Management, AI and Data Management, Dell Technologies
Use cases
ProcurementGPT
Business problem. One of the critical challenges in procurement is ensuring that contracts exchanged between organizations, such as banks and suppliers, are aligned and free from discrepancies.
Need. Efficient, reliable tools to review contracts, manage supplier relationships, and maintain regulatory compliance, where human oversight may overlook critical contractual clauses.
Parse documents into sections or clauses. The app automatically parses documents into their sections or clauses as they are loaded.
Identify the most similar sections. Retrieve the most similar sections, with their context, from the counterpart document. This works in both directions, left to right and right to left.
Break a clause down into requirements. A clause may contain several requirements, each met in a different part of the contract. Each clause is broken into its key requirements to find where each one is fulfilled, and the most similar requirements are retrieved from the counterpart document to check whether each was met.
See which requirements were met. A green tick beside a requirement means it was met; a red cross means it wasn't. The top 10 most similar requirements decide the result and show when a requirement is clicked. An overlay highlights met requirements in green and unmet ones in red, to jump straight to what's missing.
Work from either document. Everything that works left to right also works right to left, so either document can be the base for finding similar sections and requirements.
Select sections by hand. If a section wasn't parsed or broken down correctly, the user selects it manually: first from the template document, then from the supplier document. Both are broken into requirements and checked against each other.
Complaint Summarizer
Business problem. Manual processing and analysis of customer complaints can lead to missed insights and delayed resolutions. A complaint summarizer addresses this by automating sentiment analysis, identifying trends, distilling recurring themes, and prioritizing urgent issues with timely feedback.
Need. Faster grievance resolution and an effective regulatory response: faster resolution times, better first-contact resolution, and accurate complaint classification and sentiment analysis, with lower complaint volume, higher productivity and cost savings through automation.
RAG workflow. A retrieval-augmented generation workflow summarizes customer complaints across categories and products.
Complaint distribution by state. The breakdown of complaints by state, highlighting the most frequent issues such as credit cards, debt collection and loan services.
Summary of complaints and key issues. A high-level summary of the most common complaints, the top issues, and recommended actions to address them.
Actionable insights. Key topics from the complaints, with proposed recommendations for improving services.
End-to-end complaint agent. The full AI-driven process: voice-to-text conversion, complaint decisioning, and summarization with recommended actions.
Resources
- Dell and H2O whitepaper: complaint summarization on Dell AI Factory with NVIDIA
- Dell and H2O solution brief (PDF)
- Dell Technologies: AI solutions
- Dell AI Infohub
See it in action
H2O on Dell AI Factory is available in the Dell Customer Solution Center. Ask your Dell rep or the Dell AI Pursuits team for a demo.