Robotic Process Automation (RPA) continues to drive topics of conversation with our clients, and as outlined in our previous blog posts, RPA Delivers ROI with an Army of Robots Part 1 and Part 2, organizations are recognizing the role of enterprise content management (ECM) and in particular, document capture and optical character recognition (OCR) as part of their strategy.
We continue to see a convergence of content with RPA, as organizations move from proofs of concept to centers of excellence, and are evaluating the role of data capture, OCR, artificial intelligence (AI) and natural language processing (NLP). This is reinforced to the point where RPA was front and center at a recent AI Summit in San Francisco; in the more than 60 sessions conducted throughout the conference, robotic process automation (RPA) and document capture/OCR were at the forefront of use-case scenarios.
Here are some key takeaways from that event:
RPA for Improved Efficiency
RPA has rapidly become a catch-all marketing term, falling under an umbrella of a number of technologies that enable AI. It is essential to point out that RPA is not a product, but a tool for crafting long-term goals with specific technologies.
A recent survey of executives showed that improving process efficiency was rated a higher priority than reducing costs by a significant margin. Executives are clearly looking to RPA to drive digital transformation and, specifically to increase speed of go-to-market, agility in go-to-market, and business process efficiency and quality.
From this perspective, RPA has garnered much attention and support from the C-suite.
The Evolution of RPA
First-generation RPA consists of basic copy and paste bots—or macros-on-steroids—for task automation and part automation such as screen scrapping, which is not as sexy as AI or digital transformation. Nonetheless, it is still very effective and clearly serves real demand.
Some of the next-generation RPA use cases are based on robots that extend the value and automation possibilities via documents and text, such as document robots, invoice robots and contract robots.
Today document capture provides the data extraction of semi-structured documents (AP automation, for example) and unstructured documents (e.g. contracts) that is considered cognitive processing and/or NLP.
The capabilities of next-generation RPA are dependent on the rise of machine and deep learning, and the challenges of aligning these technologies with business efforts, in addition to the current state of machine language, the challenges of universal data modeling, and the role of natural language in processing text analytics.
The Next Wave: Capture 2.0
It is interesting to see RPA technology companies viewing capture as the next shiny thing and marketing it as part of the automation stack. In reality, intelligent capture has extended the overall digital transformation narrative for some time, and has focused on task automation for many years.
The core elements of intelligent capture that are driving this interest include:
1. Automatic Classification and Processing
While OCR technology allows for text recognition, today’s capture platforms take one step further by being able to read the information on that document, classify it correctly, and automate workflows based on that classification.
While the system is initially guided by a set of rules, its identification and processing capabilities continue to improve using machine learning—meaning it is able to learn from repeated exposure to documents, as well as from the actions taken by users upon those documents.
As a use case in accounts payable, intelligent capture can automate the processing of invoices by recognizing elements such as invoice numbers, line items and so forth, despite the fact that these elements can appear in different locations and in varied fonts and sizes, depending on how any given company chooses to format its invoices.
2. Data Extraction
By being able to accurately read information and understand context through leveraging AI and NLP, capture can take data extraction to a whole new level. This is an ability that is more vital than ever, as organizations are inundated with more and more data.
3. Document Clustering
Document clustering, whereby documents are grouped by topics without prior classification, is another ability made possible by AI and NLP.
This can lead to understanding how documents relate to one another within a wider context, and helping find similarities and make inferences that perhaps would not have otherwise been possible.
4. Advanced Security
AI-powered document management systems can help enhance security and protect customer data, by detecting sensitive and personal identifying information (PII) and flagging these documents for special handling. Automatic classification and processing also means documents aren’t left in unsecured locations while waiting to be actioned.
5. Data Analytics
Perhaps the most exciting prospect of AI and NLP is the potential for analytics and the value this can bring to organizations.
Incorporating capture to classify and collect data, combined with platforms, predictive analytics and data visualization, allows organizations to improve insight into business context and improve decision-making and business processes.
The “Next Shiny Thing”
Just as many of the process automation companies have discovered, if your organization is looking at ways to automate your business processes, either through BPM or RPA, maybe it is time to revisit capture and the capabilities that are being delivered through the next wave of capture 2.0.