Digital Worker DNA

Digital workers are increasingly able to act in many ways similar to humans. While they are a long way from coherent creative thoughts, they will be able to evaluate situations using available data and then take the best available course of actions for a desired outcome.

In the case of rise of digital workers, we see the market moving across many components that have begun to converge already. While no one organization has created a complete digital worker, collectively using a best of breed approach with available technologies is making progress.

What makes up a digital worker?

Illustrated in the chart below, we see three key capabilities of the digital worker. Machine learning (ML) and artificial intelligence (AI) are technologies that support and enable the digital worker’s higher processes.

Actions capabilities

Working from the bottom up, Actions is the ability to use technology as humans do and is at the core of the digital worker, whether executing a long running set of predefined activities or performing a simple process. This is the provided by robotic process automation, also known as RPA. Organizations often start with Actions as it is a clear path to early value which can be the seed funding for the new organizational capability.

Three platforms have traditionally led the market for Action capabilities:

These platforms began by providing Actions capabilities to the marketplace. In their own unique ways, they have added Awareness capabililities whether through product extensions and/or through connectors to other best-of-breed tools.

Awareness capabilities

In the middle layer is Awareness which is the ability to digitally simulate human senses and includes internet-of-things (IOT), natural language processing (NLP) and optical character recognition (OCR) among others. The Awareness capabilities structure data that is unstructured (e.g., an email) and semi-structured (e.g., invoice). Awareness greatly extends the available use cases to the Actions layer and creates much more opportunity for value. Combined, Awareness and Actions deliver what is commonly referred to as Intelligent Process Automation, also known as IPA.

There are many platforms that deliver Awareness capabilities, top providers include:

These platforms are best-of-breed tools for structuring data, bring core features such as optical character recognition (OCR) while continuing to invest in more advanced features such as natural language processing (NLP).

Alternatives capabilities

At the very top, Alternatives is the ability to identify patterns and evaluate as a human for a defined outcome or new source of value. Using the structured data to understand a situation, Alternatives seeks to optimize the outcome by determining the best path forward with available options. Thereafter, it directs the digital worker to perform these activities.

Enhanced by machine learning and artificial intelligence, this micromanagement capability will drive further use case expansion to maximize overall value for the organization.

There are many platforms that deliver Awareness capabilities, top providers include:

Unfortunately for now, the Alternatives layer is the most challenging and, consequently, is commonly managed by data scientists. As it takes a great data to train these capabilities, we recommend that early adopters license someone else’s model for common needs and limit their own model investments for their unique competitive advantage.

Toward that end, we’re seeing companies beginning to market models for common needs. For example, you could license expense account audit models to detect for fraudulent submissions and contract analyzers to detect undesirable commercial terms.

At scale, we expect the current AI market to remain model-driven rather than platform-driven for the foreseeable future.

RPA today and tomorrow

No organization has yet put together the complete digital worker as described above. But digital worker revolution has begun. Today, we are successfully deploying digital workers that could perform a significant portion of work that human operators are tasked with currently. These activities are typically those that are trainable and repeatable, meaning that they can be done the same way time after time, thereby completing the routine work, leaving the exceptions and outliers to their human cohorts.

The robots can work in nearly any technology environment, including legacy ERP systems, homegrown systems, vendor-provided interfaces or the newest cloud/SaaS solutions. For example, a software robot may take incoming PDF forms via an email account, read the form for its needed data and then populate various systems with the data fields and validation checks required and send approvals or notifications to the next step of the process. Cases that create exceptions or errors may be sent to a human operator for correction, still enabling the bulk of the workload to be handled faster and more accurately without human intervention.

Smart companies shouldn’t wait for the digital workforce to be perfected. It is necessary to learn now. Only by implementing and learning through the “simpler” automations will we be ready to handle the more advanced later. Organizations need to master the “robot way,” meaning redesigning core processes for best-of-breed for how a human and computer would perform the work. All of this will be necessary groundwork in our evolution. The case for doing intelligent robotic process automation (RPA) now is also clear.