Data analytics startup Incorta Inc. is looking to increase its relevance in the generative artificial intelligence industry with the debut of a new offering called Operational GenAI, which it says can become the foundation of intelligent business applications.

The company said Operational GenAI is private and secure, integrating state-of-the-art AI techniques such as retrieval-augmented generation or RAG, plus model serving and fine-tuning capabilities with customized on-premises hardware built by Super Micro Computer Inc.

Incorta is the creator of a cloud-based platform called Operational Lakehouse, which reduces the time it takes for businesses to make their data available for analysis. It removes the need to use extract/transform/load or ETL tools to pull data from different systems into a centralized database. It can pull information from multiple databases and systems of record, centralize it in a single repository for processing, run analyses and then visualize the results with graphs.

That platform is now being tweaked to cater to generative AI workloads, the company said. The new Operational GenAI offering integrates the company’s flagship platform with various best-of-breed open-source large language models, plus tools to fine-tune them so that companies can infuse them with their own knowledge to ensure accuracy and safety.

To do this, it’s partnering with the RAG-as-a-service startup Vectara Inc. and the no-code/low-code AI development and deployment platform aiXplain Inc., which handles RAG plus model serving and fine-tuning, respectively, directly within the Operational GenAI platform. By using Supermicro’s AI-optimized hardware, companies can keep their new generative AI data foundation in-house, ensuring privacy.

The company argues that if generative AI initiatives are going to be successful, they need access to fresh, detailed information that’s not only secure, but also private. Most enterprises lack this, which is where the new Operational GenAI service comes in. It gives companies a way to fully automate generative AI applications with access to unlimited data in a secure, customizable environment where data privacy and governance standards are always upheld.

Incarta co-founder and Chief Executive Osama Elkady said the partnerships with Vectara and aiXplain mean it is uniquely able to combine enterprise operational data with unstructured information. “This enables customers with a click of the button to have access to an unlimited, private, and fine-tuned generative AI solution on top of the best data foundation for AI,” he said.

AiXplain founder and CEO Hassan Sawaf said generative AI is evolving to become a world of “agents” that can drive innovation and operational efficiencies in hundreds of very specific niches. These agents are all based on LLMs and then trained to perform certain tasks to a high degree of excellence, but to do this, enterprises need the right platform, he stressed. He therefore likens Operational GenAI as a primary AI agent that directs the execution of other AI agents.

“With aiXplain, Vectara’s RAG and Incorta’s Operational Lakehouse platform, applications can seamlessly become AI agents directed by Operational GenAI,” he said.

Image: SiliconANGLE/Microsoft Designer