Weav.ai emerges from stealth with a series of ‘copilots’ for analyzing unstructured data
Weav.ai, a startup that develops generative artificial intelligence “copilots” that help users navigate unstructured data, today is emerging from stealth mode and launching a series of tools aimed at information discovery and classification.
The company’s Enterprise AI Copilots come in three versions for document, conversation and search, with a pre-integrated AI infrastructure stack that combines integrations, prompt management, foundation models, vector databases, security and monitoring, the company said.
Copilots can be used to extract summarized data from structured and unstructured documents such as contracts, invoices, financial statements, resumes, knowledge bases and emails using natural language search statements. Built on a low-code foundation, the assistants analyze uploaded documents and extract relevant information such as customer account numbers, orders, purchase orders, phone numbers and financial data. Users train the model by creating a data set with a few hundred or thousand examples and can apply reinforcement learning to improve performance.
“Speed to value is what we are all about,” said Peeyush Rai, founder and chief executive officer.
Rai said the technology, which is independent of an underlying large language model, can understand many document types out of the box. Complex and industry-specific document types can require a couple of weeks of training.
“We have been able to show value in two to three weeks and first implementations in four to six weeks,” he said. “You always don’t need fine-tuning if the LLMs are already good at certain concepts and domains. Fine-tuning involves creating ground truth data from a few thousand records with humans overseeing the right and wrong answers.”
Rai said the technology has received strong interest from large enterprises and the company is in pilots at five Fortune 100 companies. One contact center that sought to audit and analyze transcribed agent conversations with customers “was able to get 1% coverage with manual effort. We are helping them get to 80% to 85% coverage with close to 90% accuracy,” Rai said.
Insurance companies are testing the technology to analyze claims and verify underwriting policy compliance. A private equity firm uses copilots to analyze hundreds of megabytes of research data that goes into making investment decisions.
For each document scanned, the system can generate a summary and display data in recognized fields for structured analysis. Multiple documents can be queried with natural language prompts.
The document search copilot is available now with a free trial on Weav.ai’s website. The conversation pilot, which Rai said will be able to analyze conversations in real-time across multiple data sources, will be available within three months. The search copilot is included in the other products. Pricing has not been set.
Weav.ai raised a $5 million seed funding round last November led by Sierra Ventures Management III LLC.
A message from John Furrier, co-founder of SiliconANGLE:
Your vote of support is important to us and it helps us keep the content FREE.
One click below supports our mission to provide free, deep, and relevant content.
Join the community that includes more than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more luminaries and experts.