H2O.ai Unveils Agentic AI that Converges Generative and Predictive AI with Purpose-built SLMs

-The industry’s first multi-agent Generative AI platform to bring together the strengths of Generative AI and Predictive AI with airgapped, on-premise deployment options. -Building on the recent releases of H2O Mississippi and Danube small language models (SLMs), this platform brings together purpose-built AI solutions for enterprise scale and regulated industries.

H2O.ai, the leader in open-source Generative AI and the most accurate Predictive AI platforms, today announced the industry-first convergence of Predictive AI and Generative AI in its enterprise generative AI platform, h2oGPTe. The new agentic capabilities enable h2oGPTe Agents to seamlessly integrate H2O.ai's predictive AI models into autonomous workflows, ushering in a new era of operational efficiency and intelligent automation.


This breakthrough transforms h2oGPTe into the only end-to-end enterprise AI platform to converge Generative and Predictive AI capabilities in air-gapped, on-premise and cloud environments, ensuring both compliance and innovation. Built for industries like finance, telco, healthcare, and government, h2oGPTe's multi-agent AI system autonomously manages complex, multi-step tasks, drawing from both generative insights and predictive accuracy to enhance enterprise decision-making with transparency and control.

"Multi-agent systems are the digital workforce of tomorrow, equipped not only to act but to adapt, collaborate, and evolve," said Sri Ambati, Founder and CEO of H2O.ai. "Our pioneering work with agentic AI allows organizations to unlock the potential of converged predictive and generative intelligence—moving beyond automation to true transformation of enterprise workflows. It's about amplifying human potential and democratizing AI, so every business can achieve exponential growth."

To further illustrate, an agent can classify customer call center inquiries into over 80 categories using a fine-tuned H2O Danube model at a fraction of the cost of traditional large language models (LLMs). This system is then orchestrated with an Agentic AI framework powered by state-of-the-art LLMs to dynamically provision operators using a predictive AI agent, enabling efficient complaint resolution.

"The development and deployment of generative language models, particularly in high-stakes sectors like banking, demand a rigorous framework that balances automation of testing and evaluation with human calibration to ensure reliability and transparency," said Agus Sudjianto, Senior Vice President Risk & Technology.

Consistency and safety in Agentic AI needs rigor, continuous reinforcement, and learning. Building trust in AI agents requires rigorous testing, thorough evaluation frameworks, and transparency through open-source development.

With its rich set of features, h2oGPTe is designed to meet these needs, providing robust agent reliability, document AI capabilities, and advanced safety protocols.

Key Features of h2oGPTe:

Multimodal Agentic AI with Predictive Model Integration
h2oGPTe Agents bring autonomous task execution to your workflows, employing LLMs to perform multi-step actions such as web research, predictive modeling, database access and iterative code execution. These agents operate programmatically to reduce manual workload and streamline operations, offering continuous, autonomous performance on tasks requiring sequential logic, data science, programming, and complex decision-making. h2oGPTe Agents can create multi-page PDF documents with charts and tables and flowcharts grounded on actual data found across various data sources, or train and deploy highly predictive and explainable machine learning models by autonomously leveraging the world's best AutoML H2O Driverless AI.

Model Risk Management for Enhanced Compliance and Interpretability
Transparent Assessments with Embedding and ML-Driven Evaluators: Embedding-based metrics complemented with Natural Language Inference provide transparent, explainable and objective model assessments to enhance accountability and clarity.
Calibrated Metrics with Human Feedback: Incorporating sampling of human feedbacks calibrate automated metrics, enabling efficient and trustworthy evaluations crucial for high-stakes applications.
Robust Testing through Automated Question Generation: Automated question generation facilitates comprehensive testing to identify model vulnerabilities and improve reliability.
Rapid Diagnostics with Visual Insights: Visualizations to enable quick identification of patterns and weaknesses, supporting efficient diagnostics and model improvement.

Coding Assistant for Rapid Prototyping
h2oGPTe's Coding Assistant helps developers quickly prototype ideas by generating starter code and scaffolding for new projects. It provides basic code completion and documentation, helping teams move from concept to working prototype faster. The assistant supports common programming languages and can suggest simple optimizations during development.

Citation-Based Verification for Transparent Retrieval Augmented Generation (RAG)
State-of-the-art multimodal RAG with built-in citation support offers comprehensive traceability for AI-generated responses, with embedded document references that enhance transparency. This feature is ideal for audit-heavy sectors, ensuring each AI response is both accurate and verifiable.

Customizable Guardrails for Safe AI Deployment
Control response boundaries and safeguard sensitive information with h2oGPTe's guardrails and PII controls. Configurable safety mechanisms allow enterprises to comply with stringent policies and ethics standards, aligning AI behavior with corporate and regulatory guidelines.

Intelligent Model Routing for Optimized Performance
By dynamically routing queries to the best-suited model, h2oGPTe maximizes efficiency based on real-time assessments of cost, latency, and accuracy. Customers can decide to mix and match from a large choice of over 30 proprietary and open source LLMs. This feature ensures optimal performance without overextending resources, streamlining the deployment of Generative AI.

Document AI with Multimodal Guided JSON Generation
With built-in Document AI, h2oGPTe produces schema-driven JSON outputs for efficient processing—essential for contract summarization, compliance metrics extraction, and structured reporting. This ensures reliable, contextually grounded outputs drawn from private repositories like document archives and knowledge bases.

Multimodal Audio and Vision Analysis
h2oGPTe's Audio and Vision Models can extract structured information from audio files, images, figures, and other visual elements like flowcharts or handwritten documents. These capabilities are essential for fields where insights are often embedded in diagrams and tables, offering a new level of interpretability and insight for data-driven decision-making in visual-heavy contexts. Audio models can transcribe and translate recordings in dozens of languages. Vision models can help AI agents autonomously verify their own generated content.
H2O Agentic AI and new features are generally available in Enterprise h2oGPTe on Thursday, Nov. 21, 2024. For more information or to schedule a demo, visit https://h2o.ai/platform/enterprise-h2ogpte/

About H2O.ai

Founded in 2012, H2O.ai is at the forefront of the AI movement to democratize Generative AI. H2O.ai's open-source Generative AI and Enterprise h2oGPTe, combined with Document AI and the award-winning autoML Driverless AI, have transformed more than 20,000 global organizations, and over half of the Fortune 500, including AT&T, Commonwealth Bank of Australia, Chipotle, Singtel, Workday, Progressive Insurance, and AES.

H2O.ai partners include Dell, Deloitte, Ernst & Young (EY), PricewaterhouseCoopers (PwC), NVIDIA, Snowflake, AWS, Google Cloud Platform (GCP) and Microsoft Azure. H2O.ai's AI for Good program supports nonprofit groups, foundations, and communities in advancing education, healthcare, and environmental conservation. With a vibrant community of 2 million data scientists worldwide, H2O.ai aims to co-create valuable AI applications for all users.

H2O.ai has raised $256 million from investors, including Commonwealth Bank, Nvidia, Goldman Sachs, Wells Fargo, Capital One, Nexus Ventures and New York Life.

Featured Product

OnLogic Helix 511 Fanless Intel 12th Gen Edge Computer

OnLogic Helix 511 Fanless Intel 12th Gen Edge Computer

OnLogic's Helix 511 Fanless Edge computer delivers ultra-reliable, fanless computing using Intel® 12th Generation performance hybrid processing. The Helix 511 is a versatile fanless computer capable of powering solutions including advanced automation, light detection and ranging (LiDAR), access control & building automation, or virtually any other IoT or edge gateway functionality needed, with support for 4 simultaneous serial connections. The system is able to reliably operate in temperatures ranging from 0 to 50°C, can accept power input ranging from 12 to 24 Volts, and is Wall, VESA and DIN rail mountable.