Enterprise software company Elastic has entered the generative AI space with a new platform called the Elasticsearch Relevance Engine (ESRE). The company claims it will allow developers to build apps that can securely link services like ChatGPT to private enterprise data.
ESRE works on both structured and unstructured data and tools built with the platform will allow end-users to query company information using natural language prompts. It can be integrated with a range of transformer AI models including those built by Elastic, or third-party tools such as GPT-4 from OpenAI.
Large language models are being used by Google and Bing to refine search results for internet users, and the same process is happening in the enterprise search space. Companies such as IBM, SalesForce and SambaNova are building platforms, systems, tools and models that provide a secure mechanism for training private and sensitive company data on these powerful tools.
Models like GPT-4 are trained only on publicly available data. This means it has limited use for businesses that would need it to be able to provide insights into an organisation’s proprietary software or internal documentation. This is something Microsoft is attempting to change with Copilot for Work, but that is being rolled out slowly and is restricted to Microsoft tools.
Elastic says its product is available now and will allow for a range of applications to be built on top of company data. Its focus is on a new search paradigm, making it easier to find and query data in a way not possible in previous iterations of AI tools.
Elastic’s ESRE powered by a range of models
As well as integrating models like GPT-4 or Claude from Anthropic, Elastic says ESRE is powered by its own built-in vector search and transformer models. “It enables companies to achieve breakthrough results by helping organisations to create secure deployments to take advantage of all their proprietary structured and unstructured data while optimising infrastructure and using talent resources more efficiently,” the company enthused in a statement.
“Generative AI is a revolutionary moment in technology and the companies that get it right, fast, are tomorrow’s leaders,” said Ash Kulkarni, CEO of Elastic. “The Elasticsearch Relevance Engine is available today, and we’ve already done the hard work of making it easier for companies to do generative AI right.”
The new platform includes tools for vector search, BM25f search, hybrid search and a new transformer model that can fit on a laptop which means the power of large language models can be used “on device”. This means “businesses and teams are now able to optimise infrastructure and talent resources more efficiently,” Elastic explained.
“Enterprises are excited about the potential for generative AI in their applications and workflows, but are all also swamped by the pace of innovation in the field,” said James Governor, co-founder of analyst company RedMonk. “ESRE is designed to ease adoption of transformers, homemade and third-party LLM models, building on the original core strengths of Elastic in search.”
Read more: IBM announces watsonx enterprise AI platform
Homepage image by Gorenkedoff/Shutterstock