Innovation and Automation: An End-to-End Approach

When we tackle automation challenges, we tend to break them into pieces so we can focus on one thing at a time. We highlight what we are responsible for individually, perhaps what is measured in our KPIs or management objectives. We talk about “cutting the problem down to size.

Sometimes this works just fine. But sometimes we don’t get the best result in automation because, in focusing on the components, we don’t take an overall view. When we break things down and sort them into familiar categories, we can fall into the trap of applying the automation tool we’re most familiar with — even if it’s not the best tool for the job at hand. And at the point when we’ve applied the tool we know best, it can be tempting to declare “job done” and move on to the next item in the work queue. Instead, we need to step back and ask whether the automation solution we’ve applied is optimal, or merely adequate.

I’d like to propose a different, more innovative approach: a closed-loop or end-to-end methodology for automation focused on streamlining operations, simplifying how work gets done and accelerating better outcomes, all with a view toward continuous improvement.

The first step involves discovery and documentation of your processes. This should occur well before selecting or applying an automation technology. The goal at this stage should be to understand how your company really works and who is responsible for critical functions. You need to know the implications and consequences that occur when things go wrong or don’t get done at all.

Most organizations are flying blind in this regard. They don’t have a good grasp either of their processes overall, or of the details of individual processes. True, there might be a handful of people in the enterprise who truly understand each process. And some companies are beginning to utilize an emerging technology, process mining, that looks at system event logs to gain deep process visibility. But process mining alone doesn’t address the all-important question of how processes should run.

In any case, few organizations task anyone with taking an overall view of processes and how they interrelate. Nor do they have a mechanism for establishing consensus on what they want their processes to accomplish, or what constitutes an optimal process. There could be many metrics for what is optimal, such as a maximum number of steps or a maximum time it should take to execute. But very few companies have even thought about their processes in this way, let alone attempted to codify their thinking and create a shared understanding.

More organizations are beginning to undertake this type of systematic process discovery, documentation and management, however. I expect it will become more common, because it often pays dividends early on. As people and teams document their processes and share their findings, they often notice overlaps, gaps, redundancies, and inefficiencies that can be addressed at a policy level, even before applying automation technologies.

Not every process is ripe for software automation, of course. If a fire breaks out in a theater, you shouldn’t open an application and kick off a workflow, you should sound an alarm and call 9-1-1. But it’s definitely good practice to document what you want to happen in case of fire.

And a large share of enterprise processes cry out for automation. Once you have completed the discovery and documentation phase, the next step is to take those results and select the right type of automation. Keep in mind the great variety that exists in both process types and automation technologies, and select automation tools based on fit, not familiarity.

For example, some processes are document-centric, and are best addressed with technologies tailored for that purpose. These solutions can generate documents by pulling information from multiple sources, and they make it easy to embed electronic signatures.

Other types of processes don’t involve documents but require interaction with a variety of applications built with APIs. These types of processes call for automation with modern workflow solutions. This is especially true if they involve collaboration, evaluation, judgment, decisions, and complex logic — that is, if each step has several possible actions and the process flows differently depending on which action is taken.

On the other hand, many enterprises still run mission-critical legacy applications that predate the emergence of APIs. Robotic process automation is seeing a surge of interest today precisely because these older applications are still in wide use, especially in industries such as banking, insurance, and health care. Often these processes are highly repetitive and manual. Many involve forms, sometimes electronic but often scanned and processed using optical character recognition technology to identify relevant information.

Thorough documentation and an awareness of the full spectrum of available solutions are the keys to selecting the right type of automation for a particular process. In addition, enterprises must be careful as they evaluate solution providers, keeping in mind that vendors who specialize in one type of automation will naturally tend to frame challenges in ways that play to their strengths.

Enterprises also must bear in mind that many processes include components that call for different types of automation. For example, preparing and executing a sales contract might start with generating a document from multiple sources; then call for a robotic element to retrieve and enter data into specific fields; then a workflow to make sure the right people review and sign off; then an e-signature component when the deal is closed; and finally another robotic element to update various systems with details such as the deal size, key contacts, discounts, commissions to be paid, and the like.

Clearly it wouldn’t make sense — and might not even be possible — to apply the same type of automation technology to each of these steps. But if your staff, or your vendor, has limited expertise, you might wind up with a sub-optimal solution whose components aren’t well matched to their steps in the process. So it’s important to ensure that your automation solutions can address the full range of challenges you face. And keep in mind that if the overall process you need to automate has disparate components — API-enabled workflows, RPA botflows, and document generation, for example — your solution must be able to orchestrate or choreograph how all the components work together.

Next, once you have applied an appropriate automation solution, it’s natural to think the project is finished. But if you stop there, you pass up an important opportunity: the chance to optimize your processes by analyzing how they are running, identifying bottlenecks and breakdowns, and modifying them to run as smoothly and quickly as possible.

To do this well, it helps if your solution supports instrumentation so you can harvest metrics on when, where, how often, and how quickly your processes run, and the person, role, or step where they bog down. Gathering this type of insight is valuable in two ways. First, it allows you to optimize the specific process in question. Second, you can take the lessons you learn from optimizing this particular process and apply them elsewhere. For example, you can feed them back into the discovery/document phase of your automation process, and every subsequent process will benefit.

These three steps — discover/document, automate, optimize — make up an end-to-end or closed-loop approach that allows you to optimize processes across your organization, systematically and comprehensively. Adopt this approach, and you’ll be at the cutting edge of automation. And remember that every process should be documented and optimized, even the ones that don’t lend themselves to software automation.

In addition to this conceptual framework, I would also like to share some quick practical tips based on my years of experience in this space.

The first tip is to start simple. As you discover and document your processes, certain ones will jump out as obvious places to start – clear problems with relatively straightforward solutions. Addressing these first will allow you to provide value quickly and build expertise that you can apply as you scale up your process optimization initiatives.

The second tip is to be agile. Don’t let the perfect be the enemy of the good. Get started, get feedback, and adapt accordingly.

The third tip is to build a team of people committed to optimizing processes across your organization. Create centers of excellence where you share experience and best practices. Recruit people with the expertise to achieve outstanding results and the passion to inspire others to join them. Grant them autonomy to move ahead on their own, but build in accountability, because without it you risk anarchy.

If you follow these steps – take an end-to-end approach that spans discovery/documentation, automation, and optimization, and then seek to create value quickly — you will be at the leading edge of innovation in automation. You’ll streamline processes across your organization and set a course for continuous improvement.