AI-Enabled Hyperautomation for Document Processing
Hyperautomation deals with the application of advanced technologies, including Artificial Intelligence (AI) and Machine Learning (ML), to increasingly automate processes and augment humans. In this article, we’ll explore how hyperautomation influences document processing.
AI-Enabled Hyperautomation Use Case
Hyperautomation extends across a wide range of tools that can be automated, but also refers to the sophistication of the automation (i.e. – discover, analyze, design, automate, measure, monitor, and reassess). The main idea behind this process is to train an Artificial Intelligence (AI) enabled document processing engine to read documents and extract data to be used by downstream processes.
During a recent Smartbridge hackathon, our team aimed to automate the process of extracting information from contracts or statements of work executed between vendors and customers, and load that information into multiple systems.
The contracts we were dealing with for this use case were statements of work associated with professional services engagement. The goal was to get the following pieces of information from these statements, and populate multiple systems like NetSuite, Microsoft Teams and M-Files:
- Project name
- Customer name and address
- Resources staffed on the project
- Billing frequency and invoice addresses
Are you ready for the automation wave?
We have explored and used various tools to automate this process. The following are details of the tools used, and highlights of the project.
IQ Bot by Automation Anywhere (AI Engine)
Automation Anywhere is a popular vendor in the RPA market today, offering powerful and user-friendly RPA capabilities to automate any complex tasks. It is one of the “revolutionary” technologies that changes the way enterprises operate. The IQ Bot possesses the ability to process unstructured or semi-structured data, and extract the information into a CSV file for further use. Most importantly, the IQ Bot can be trained to extract the data based on unique user needs
During this hackathon, we were able to train the IQ Bot to read 20 semi-structured contracts after a bulk upload. We noticed that it took the IQ Bot 5-hours to process each document. Once the documents were processed, we had to manually map the information to be extracted into a CSV.
UiPath Taxonomy Manager (Process Contracts)
UiPath is another leading RPA vendor, providing a complete software platform to help organizations efficiently automate business processes. The UiPath Taxonomy Manager is a document type manager which is used to categorize the documents.
Digitize Document is a prebuilt activity in UiPath which is part of UiPath Intelligent OCR Activities. This pre-built activity can be used to extract data from the Taxonomy Manager by digitizing the document and creating a template. The Position-Based Extractor can be used to define Data Extraction Scope to extract specific content from a document. In this use case, we decided to use UiPath’s AI-based document processing engine.
High-level document of the hackathon use case
Create the Project in NetSuite – The bot consumes the CSV file exported by document processing in the AI engine, and creates a project in NetSuite while assigning resources to the API.
Email Communication – The bot identifies the list of resources and project team based on the form, and sends communication to the team with an attached SOW (Statement of Work).
Create a Communication Channel in Teams – The bot logs in to Microsoft Teams and creates the appropriate folders based on the project name, and adds the project team to the channel for collaboration.
M-Files – The bot then creates the appropriate folders in M-Files, and stores the SOW and project forms for future reference.
Hyperautomation Use Case Conclusion
It was a learning curve for us to train the two AI-based document processing engines, and extract all the data required by the downstream processes. In the end, we were able to extract all the information from the different types of contracts, other than the resources staffed on the project.
We ran into a challenge with extracting the resources because this information was not consistently available in the same format in all the contracts. Using the information that was extracted, we were successfully able to create a project in NetSuite, but could not assign resources to the project. We were able to create a Teams site and the required project folders in M-Files. We are continuing our study on hyperautomation and exploring more tools to automate document processing in the future!
There’s more to explore at Smartbridge.com!
Sign up to be notified when we publish articles, news, videos and more!