Organizations use paper. Businesses in every vertical have always relied upon a certain quantity of paper since the dawn of time and, despite the rise of digital, that basic truth isn’t going to change. Although invoices are now largely digital, boarding passes live on smartphones and there’s even an app (or two) for To-Do lists and notepads, a certain amount of paper will always feature as part of any firm’s operational tasks.
While Western Civilization replaces some of its paper consumption with technology, business growth in developing nations and the use of paper in packaging is thought to offset the trend for digitization.
Never been ‘inked’
If we accept these basic propositions, then it should logically follow that the technology industry seeks to provide solutions to working with data in its digital form. What was once printed paper, is now being digitized through OCR (Optical Character Recognition) and document capturing technologies. What was once spooled off of some form of printer, is now ‘digitally native’ data that is potentially never exposed to the wonders of ink, because its digital form has been validated for whatever enterprise use case it needs to perform.
With its roots in Eastern Europe and a foundation in document capture and management technology, Abbyy positions itself as something rather more than a content capture specialist. The company acquired Philadelphia, Pennsylvania-based TimelinePI back in May 2019 for its process intelligence competency. Add document management to business process analytics and you get a method for visually modelling data and events as they flow through a business. Now a product rather than a stand alone organization, Abbyy Timeline is a key part of what the company calls Digital IQ.
“When we digitize successfully, we (as businesses) start to find that traditional use cases of technology (that might have previously only been touched by the IT department) start to become ‘subsumed’ into other line of business applications, which is why we talk about ‘citizen developers’ now creating new application services because they can more directly ‘shape’ the way powerful back end technologies function,” added Macciola.
Once documents (or images) have been scanned and digitized, then metadata can be applied to the stock of information that exists inside an organization using this technology. Metadata is sometimes described as ‘information about information’ and it works to classify data into various groups and schema, which allows it to be more accurately stored in data stores.
Information ‘Pathways’ Denote Processes
A typical company might have several hundred different ‘paths’ down which data might flow within the business. These pathways denote, describe and define the shape of the processes inside the organization itself. Abbyy Timeline works to define a set of rules that a customer can use to look inside any single ‘process instance’ and see how work is actually executed. Some process workflows are repeatable and easier to label, some are ad hoc and more random in nature and some are in between the two i.e. not massively repeated, but still somewhat variable. Abbyy Timeline ingests all the event data inside these work processes and then builds a visual model (on screen) to identify bottlenecks, where work may still happen, but not in a way that is as fluid as it could be.
Once a business starts to get all the records, documents, events, images and other operational data appropriately digitized for more automated business intelligence, it will typically need to clear out anomalies which stand out. On the road from analogue to digital, some data gets missed, some data isn’t formatted as it should be and some data gets badly scanned.
Over and above data ingestion and the chance to produce nice on-screen images and tools to explore that data with, Abbyy also focuses on the validation of the business data and the classification of the data. The company says that these stages need to be executed before the onward management and analysis of operational information. Company CEO Ulf Persson explains that all companies know that they have all these processes and they know that they have desired outcomes from those processes but that (most companies) fail to understand how complex they are in terms of how many steps, stages, people, products and things that are involved.
«Whether you look at your data using enterprise reporting, business intelligence dashboards or even cool tools like geospatial visualizations you are still limited to only looking at a single step of a business process at a time. With the advent of first-generation process mining software this improved as it allowed you to visualize all steps of a business process end-to-end – as long as the process was very well defined and consistently executed,» said Scott Opitz, president of Abbyy Timeline. «The problem is, when you are dealing with more convoluted or complex business scenarios where the steps of the process are performed in a more ad hoc manner such as a hospital emergency room or a wide-ranging customer service situation, where you can’t always resolve the process to a nice neat diagram. In these situations you need a new approach to analyze the process which is based on the complete process timeline – not some mythical pretty process diagram.»
Opitz uses this proposition to attempt to validate and justify the need for the types of process tools that his company specializes in. Going further, Opitz says that Abbyy Timeline has also implemented machine learning techniques to predict process outcomes and prescribe actions based upon user-defined priorities. So in the real world, this means that the system ‘maps’ the way processes are carried out and accommodates for the fact that different employees or contractors will execute any given process performing steps in different orders and combinations, at different speeds, at different costs, at different levels of competency — and the software ‘learns’ each time a process is digitally mapped to be able to make the system cumulatively smarter over time.
Alongside its Timeline product, Abbyy also points to its Vantage tool. This software offers human-like cognitive skills to automation platforms such as RPA (Robotic Process Automation) and BPA (Business Process Automation). The technology delivers text recognition, machine learning in the form of ‘consumable skills’ for non-technical workers. The company insists that Vantage addresses common content-centric pain points such as lengthy set-up and configuration, shortage of training data, requiring skilled experts in areas like machine learning for automating unstructured content.
Digital Business Codification
Where all this takes us to is digital enterprises. The marketing people love to call it ‘digital transformation’, but it might be more accurate to call it digital codification i.e. we’re taking existing elements of business and driving them into a place where we have coded the information in the processes that drive them.
Some parts of business are horizontal (in that they apply to every industry vertical) such as invoicing or facilities management. Some parts of business are vertical (in that they have a specific shape, role and context within a specific type of business) such as insurance claims management, cake baking or root canal treatment. All of these parts of business form a process and so we can digitize the paper-based documents that used to accompany them.
Even when a business process is ‘very human’ and doesn’t have a defined shape or piece of paper attached to it (such as a worker going to a different department and having an informal chat with another team), we can usually ‘infer’ that something has happened in a work process without knowing its exact shape based upon later results.
Paper isn’t dead, but work has gone digital… and you can make a note of that!