Alkymi

Alkymi is a New York City based developer of an enterprise platform designed to help financial services companies automate manual tasks on documents. Alkymi markets its product towards companies in capital markets, or those operating in the banking, asset management, wealth management and alternative assets sectors.

Alkymi's platform works to understand a clients data and workflow using deep learning and computer vision paired with human expertise. Alkymi breaks this process down into three unique stages.

In Stage 1, Alkymi's platform ingests and normalizes a range of unstructured data such as douments, PDGs, images and XML. The pipelines use image recognition and natural language processing with unique domain knowledge to develop document understanding.

In Stage 2, Alkymi's platform works to understand the user. This includes pushing documents with high confidence scores straight to extractions while documents with low confidences scores are reviewed. Through the review process, Alkymi learns what clients need from documents and improves their deep learning models.

In Stage 3, Alkymi uses the understanding of the clients need to automate greater amounts of document workflow through which their platform extracts the relationships, entities and concepts and delivers them to the user in formats easy to integrate with existing systems. The structured data allows users flexible analytics and customized searches for increased data discovery and content metrics.