The automation of business processes has been a common business goal for the past decade. Implementation has come a long way, moving from projects focused on reducing the cost involved in back-office processes to improving the efficiency of front-office processes responding to the ever-increasing expectations for customer experience. Yet there remains significant opportunity.
Analyst firm Infosource has analyzed the intelligent information capture market for several decades. It has reported on the increase in front-office applications that have driven the growth of the capture market prior to and during the pandemic. Going forward, Infosource expects capture to be integrated even more tightly with customer facing B2C transactional processes as the shift to cloud and the increasing incorporation of AI, and machine-learning based technology transforms the capture market.
Advanced capture is a step change for the automation of unstructured business inputs
Unstructured business inputs have remained a major hurdle for the automation of business transactions. Most inputs for business transactions are unstructured, and the volume continues to grow exponentially. In an industry survey conducted by AIIM, respondents reported an average of 57% of their information is unstructured — emails, documents, text messages, etc. The respondents expect a continued growth of incoming information; on average, they expect their current volume of incoming information to increase by a factor of 4.5. While not all incoming information is relevant for business transactions, there is a clear trend for increasing digital inputs. See Figure 1 below.
The automation of workflows involving semi-structured documents, with invoices being the most common across all industries, has been an important step forward. Invoice processing continues to be a large and increasing target market for capture vendors. Intelligent capture has evolved to handle nested tables that are common to line-item extraction across accounts payable and accounts receivable documents and enabled their broader adoption. Recently, robotic process automation (RPA) vendors with intelligent document process (IDP) capabilities have identified this large and still growing opportunity for their solutions.
The ingestion of unstructured inputs in critical business processes, like the onboarding of new customers, can involve many different types of documents combined with other data that are received through different channels and in different formats. See Figure 2 below.
The integration of artificial intelligence-based capabilities in information capture solutions marked a step change for the automated ingestion of unstructured inputs. Leading capture software vendors have added features like intelligent classification and machine learning to their established recognition and extraction capabilities. In addition, a new breed of vendors entered the capture market and applied their AI technology to unstructured information inputs.
Intelligent capture and RPA are at an early stage of maturity
While advanced capture solutions have advanced the automated ingestion of unstructured business inputs, the market is far from mature. Infosource estimates the overall global opportunity for information capture is $34 billion, with less than 20% of this total having been addressed. Infosource considers part of the current capture market to be at level 2 of the majority curve, with a larger portion between Basic and Advanced Capture, and a small portion to be further advanced. See Figure 3 below.
Many business processes are partially automated, and still involve manual process steps for approvals, validation, and manual transfer of data from or to legacy systems. The COVID-19 pandemic has exposed these manual steps as hurdles for mission critical workflows during lockdown periods. The pandemic also surfaced a high correlation between business success and the level of digitalization. Success criteria include the level of end-to-end automation, ability to ingest multi-channel inputs and systems that support remote work of knowledge workers.
A recent survey conducted by AIIM validates that organizations consider their competencies in RPA particularly low, machine learning and intelligent capture capabilities scored just slightly better. See Figure 4 below.
RPA is moving from hype to real business value
The fact that many well-trained employees spend a significant portion of their time on repetitive, routine tasks fostered the emerging market of RPA. It started with basic use cases involving screen scraping — extracting data from websites and screens of legacy systems. Over the past four years, this technology segment has experienced explosive growth supported by significant funding provided by VC firms, the backing of leading SI firms, and in some cases overstated end customer expectations.
RPA solutions have addressed the important step of analyzing and documenting manual tasks and associated business processes. Bots record keystrokes of knowledge workers, which identifies processes with significant manual steps and presents key opportunities for automation. This is very valuable for organizations, as previously this foundational effort involved time-consuming and cost-intensive process analysis. The applications range from use cases that bridge siloed business repositories like business analytics and compliance reporting, to the automation of internal workflows like expense reporting, to external transactions like invoice processing and employee onboarding.
The global RPA market grew at a compound annual growth rate of over 100% from 2016 to 2019. In 2020, despite the challenges during lockdown periods, the RPA market increased by over 50% to $2.3 billion. Infosource estimates the capture-related portion of the RPA market to account for $440M in 2020, with half of this being IDP applications. See Figure 5.
Primary use cases for traditional capture software applications overlap with IDP solutions
The examples of use cases in Figure 6 indicate whether they present established applications for traditional capture and RPA solutions. It illustrates the overlap of classic intelligent capture use cases like invoice processing and customer onboarding with IDP type RPA solutions.
Business objectives related to information capture have started to shift from a focus on cost savings in back-office operations to improving the customer experience in front-office workflows. This has driven an increasing demand to automate case management applications, which both traditional capture software vendors and RPA vendors have fulfilled. There is a broad range of case management-type use cases including customer/patient/employee onboarding and claims/mortgage applications, which present a significant future potential.
In addition to IDP type solutions, major RPA target markets include customer support applications, internal workflow automation like IT operations management,
and reporting-focused use cases such as journal entries. See Figure 7 below.
In 2020, the two major application groups in the traditional capture market were accounting and case management, each representing about a third of the global market value. In the capture-related RPA technology segment, IDP use cases accounted for over half of the end customer market. Case management applications represent the largest IDP application. Customer experience management (CEM) related applications account for almost a quarter of the capture-related RPA market segment.
Expected convergence of the RPA and the traditional capture markets
As RPA vendors increasingly target business applications that involve unstructured business inputs, they recognize the need for capture technology. For their IDP solutions, they have partnered with capture vendors or licensed their technology. Some have even developed their own capture. Capture vendors have identified the opportunity for expanding their workflow capabilities and acquired RPA startup companies, established partnerships with major RPA vendors or developed RPA technologies themselves.
Infosource expects the traditional information capture market and the capture-related RPA market to converge as traditional capture solutions and RPA solutions often target the same use cases. These opportunities will either be fulfilled by capture vendors through their expanded portfolios or RPA vendors providing IDP type offers or partnerships with capture vendors.
Organizations should carefully analyze the level of automation of critical business workflows that have a material impact on the success of their organization. Identifying hurdles and challenges identified during the pandemic should augment this analysis.
When assessing end-to-end workflows, like order to cash or an entire claims process, this analysis will likely identify partially automated process steps, or disconnected processes for different input channels. A priority list for addressing the most critical gaps should be established as well as a list of selection criteria.
Where the business workflow involves a large number of manual steps that absorb valuable time of knowledge workers, RPA capabilities should be considered. For workflows that involve document intensive business transactions with unstructured inputs, advanced capture capabilities should be a critical selection criterion.
There are many vendors to choose from, including traditional capture vendors, RPA vendors and specialized providers of process automation. As partnerships and mergers and acquisitions continue to progress, an increasing number of vendors will offer multiple capabilities. It is key to evaluate proven installations with similar requirements and/or go through a POC phase to ensure that the solution aligns with specific business requirements.