Alkymi kicked off a packed New York Enterprise Technology (NYETM) Meetup with a demo of how the technology understands documents to automate manual workflows in financial services. The monthly NYETM MeetUp at Work-Bench has more than 5,000 members, going strong since 2012.
Key takeaways for next-generation robotic process automation (RPA)
In the presentation, Alkymi CEO, Harald Collet, and co-founder Adam Kleczewski talked about this "gnarly document problem" of extracting valuable data from emails and documents—a process that is still entirely manual for many financial institutions. Without unstructured data processing, many RPA solutions can't be effectively utilized to speed up other processes.
Key takeaways include:
- 50% of financial services workflows slow down when customer data inside emails and documents has to be manually processed.
- When you look at a financial document, there are a number of visual cues—tables and charts—that lead human processors to the information to be extracted.
- Similarly, as emails and documents stream in via API, Alkymi Data Inbox employs document understanding—computer vision, NLP, and machine learning—to extract critical data.
- Once this data is inside the Alkymi UI, a human-in-the-loop (HITL) can easily review what's been collected, validate the information, and the data gets sent downstream via API to any platform, system, or worfklow.
- Active learning via HITL assures that the machine-learning models get smarter and improve over time.
See the full presentation here. Or contact Alkymi to learn more.