GigaSpaces, an IBM Gold Partner, is taking its collaboration with the tech giant one step further: The two companies are developing a solution based on WatsonX that will enable organizations to run Generative AI tools to query organizational data in natural language – and in significantly higher accuracy than today’s industry standards.

This solution, named eRAG (enterprise retrieval augmented generation), is based on adding context to structured data stored across companies’ databases. This contextual data enables LLMs to “make sense” out of structured data, which in turn improves the accuracy of the responses that GenAI can provide on questions concerning this data.

This eRAG solution will be unveiled at TechExchange: IBM’s premier annual technology conference for clients and partners, held on Jan. 22-25 in Barcelona.

“GenAI is by far the biggest promise in technology today, and almost all big tech companies are in a race to implement natural language interface onto their products”, says GigaSpaces CEO Adi Paz. “The challenge that these companies are facing is in bridging the gap between LLMs and the actual databases that store enterprise data – which is typically organized and cataloged in a complex and unintuitive fashion. That’s precisely the reason that so far most attempts to implement GenAI on structured data have very low accuracy rates, which yield irrelevant responses to queries or even worse: completely wrong answers.”

Mr. Paz will introduce this solution on the center stage at TechExchange next week, in a panel that will include senior executives from AWS and NTT Data.