Why Bother With Robots?

As a former IM and records consultant, I have worked with a lot of organizations looking to be future forward and move to a fully digital way of working — their “digital transformation.” In my experience organizations successfully moving to fully digital have some overarching strategic drivers as part of a multi-year plan.

Your move to a digital enterprise needs an information extraction and use strategy. There are three main drivers of a successful digital transformation (DX), in my experience:

  1. Focuses on the business’ future needs. This might seem obvious, but successful DX strategies work backward from a future state to define technology that will provide platforms of the future, not merely bridge the gap.
  2. Instills process flexibility through streamlining. There two levels to this. One is streamlining user requirements for a process; i.e., automation of tedious steps. The second is ensuring information can move in “quantums” — the idea of information being used simultaneously in multiple processes.
  3. Acknowledges, but doesn’t fixate on backward compatibility. You can’t make all of your vendors, partners and customers magically be “digital,” but you can make it easier for them to use your digital methods of communicating than to stick to the old methods. Successful DX strategies have a focus on ingestion and routing of information that is content-type agnostic.

Mapping software to digital transformation

If you do one thing to prevent being a partial failure (a tough prospect as roughly 84% of digital transformation projects fail), focus on ingesting and extracting information from sources at the first point of contact. The root cause of DX failure is the assumption that identification and organization of information can happen at any time. It never happens — garbage in, garbage out is still the rule.

Wading through the miasma of advice

Let me start by saying I hate buzzword bingo. It is the bane of the information management field. It distracts from solving problems by inflating the value of a specific piece of software with an unrealistic view of the usefulness of a particular piece of software, to say nothing of the implications for your IM strategy that is required to properly implement said software. [Author’s note: Irony is acknowledged]

Robotic Process Automation (RPA) is the current buzzword category that aggravates me. RPA projects are no more successful than any other digital transformation technology. This means that while RPA is sexy, it is not effective.

Why?

The TL;DR version is this: RPA is not a future-forward platform. It solves a gap in your user experience strategy with inflexible screenscraping and click-tracking. This limitation is why so many RPA vendors are partnering with AI and capture vendors — so why bother with RPA if it needs two other products that you likely have or are considering?

So how does RPA fit (or not)?

RPA is by far the best technology for backward compatibility and automation of tedious tasks. It is primarily applied to labor-intensive manual processes, such as capturing data from third-party systems. Using a combination of screen scraping and “click monitoring” it can mimic a user’s extraction of information from scanned forms with what is essentially a bot, or “robot.” This is especially useful in markets like financial services, where capturing standardized documents quickly has a measurable value.

The issue is that it cannot capture what the user doesn’t value, or if the format of the document changes. RPA, by itself, cannot expand the uses of information; it can literally only take tedious tasks away. It is great for upping user productivity today but it lacks the flexibility and future forwardness to be a platform for digital transformation.

So is there a better plan for digital transformation?

I will fully admit I am biased — the successes that I have seen (and driven) have been focused around separating information from the artifact. By that, I mean we need to extract the information, identify it based on internal practices and language to make it available to all users regardless of what system they work in, and bring the information to them — don’t make users fetch it.

So what does that mean for digital transformation strategy?

For me, I don’t see why you need RPA

Today’s capture platforms have the capabilities to extract all of the information from any artifact, paper or digital, classify the information type based on the process that it is part of, and organize it for use in multiple systems. The information is then sent directly to the users in their workstream. This is the table stakes of what the industry calls intelligent/smart/cognitive capture.

What capture lacks is the flexibility to deal with human syntax to understand all the information conveyed in a customer email or tweet or a structureless invoice (e.g., an email), which is the form of content that is causing the most pain for customers. It is also not a content type that RPA can manage due the lack of structure.

In an intelligent capture system, this would be a human task, but if we augment an intelligent capture solution with AI, AI can be leveraged in a way that expands capture into these novel unstructured content types.  As a bonus, it helps show ROI for the adoption of AI as business software, not a cute pet project.

The type of AI that is deployed to augment the capture system’s is a full AI system with true language processing, understanding of “feelings” of the author, factual statements versus opinion that allows for automation of a process intelligently — “IPA” if you will. Since the combined AI and capture can use your company’s unique terms of reference and extrapolate content in a captured document based on context, new documents can be routed without retraining the system and ensure that the user sees that system in the context of the relevant work task. This is the true foundation of digital transformation; flexible, foundational technology that provides information in a way that can be automated and presented where users need it.