By Sébastien Labourdette and Nick Daley
We are very excited to be backing Hypatos, a technology company enabling large enterprises to automate complex back-office operations by combining deep learning-based machine vision with natural language processing (NLP). Plug and Play Ventures is glad to participate in Hypatos’ €10M Seed round, alongside with UVC Partners, Blackfin Capital Partners and Grazia Equity.
Most back-office tasks still rely heavily on people and paper generating annual cost of more than €2tn per year. Manual document processing is expensive, slow, error-prone, and demotivating. Yet today’s technology – Robotic Process Automation (RPA) – is limited in scope and performance. RPA companies require structured, consistent inputs and apply rule-based automation. But they lack the document and context understanding piece, and are therefore limited in the use cases they can work on.
This narrow scope has given way to the emergence of deep-learning-based approaches to automating financial document processing. These approaches are centered around providing an understanding of documents as a first step prior to automation, therefore enabling the processing of more complex documents, and sometimes more complex decision-making. Eventually, the introduction of AI into the RPA technology stacks is recognized by the industry as the necessary step prior to true mass-market adoption of automation, in addition to ease of implementation.
But why aren’t existing actors picking up quickly on CPA themselves? The complexity of developing such technologies increases with the number of use cases that you want to cover. Hence, technological complexity meets the innovators’ dilemma, unless the company manages to generalize its machine learning models to broaden potential applications, or reduce training time to adapt new models. On the other hand, some innovative, young companies are bringing CPA to the market but try to overthrow established RPA vendors, who have grown very quickly & are well entrenched in their market.
The Next Era of Automation for Back Office Tasks
Leveraging 10M+ hand-labeled entities, Hypatos’ technology combines deep learning-based machine vision with natural language processing (NLP), and they outperform industry standards by >10x when it comes to understanding information on certain types of documents. Hypatos currently focuses on specific processes: Accounts payable, Order-to-cash, Tax, and Travel expenses. Their technology can handle semi-structured and varying inputs, and outputs end to end document automation of 40-50% of back-office processes. Hypatos automates complex document validations and enrichment tasks, and seamlessly integrates into existing RPA infrastructure, therefore uniquely leveraging existing automation players to enter the market.
Hypatos is dramatically decreasing the cost of document processing for back-office management, while simplifying integration and broadening adaptability of existing ML models. As a result, enterprises can automate costly processes at scale, without disrupting their existing RPA infrastructures in place. Hypatos understands semi-structured documents by combining two machine learning technologies together to understand meaning, inferences between words, and context.
At Plug and Play, we are very impressed with Hypatos’ technological achievements and their approach to the market. Both are already paying dividends, with Hypatos already showing impressive traction in DACH and Europe and signing up anchor clients in the US and Japan. More and more international enterprises are trusting Hypatos to complement their existing RPA efforts. Another validation from the industry is that not only is Hypatos convincing enterprises to implement its technology, but also technology integrators and auditing companies are taking the initiative to push Hypatos’ solution to their own clients.
Lastly, we are excited to partner with a very senior, complementary team with experience in Management Consulting, Venture Capital and Venture Building, Corporate IT integration, AI development and integration consulting, and senior management of Data Science teams at high growth companies. Dr. Uli Erxleben has held senior positions at McKinsey in Berlin and Palo Alto, and also led Rocket Internet’s growth in North America. Cem Dilmegani also held senior positions at McKinsey before overseeing technological integrations at a regional telecom operator, and later on founding the AI expert website AIMultiple. Dr. He Zhang was Head of Data Science at Lesara, and prior was Lead Data Scientist at the delivery giant Hello Fresh, supporting its massive growth until IPO.
We share the team’s ambitious, technology-driven vision and are excited to welcome Hypatos to Plug and Play’s family. We can’t wait to work with them to become a global leader in the future!
Plug and Play is one of the most active VCs in Silicon Valley, our goal is to fund the teams that are building the defensible businesses of the future. Join our ecosystem today.