Sandy Kemsley’s Vlog - Best practices in business automation application development - design #2
Sandy Kemsley Photo

Business automation best practices

#3 – Application development – Design (part 2)

By Sandy Kemsley

Video Time: 8 Minutes

Hi, I’m Sandy Kemsley of I’m here for the Trisotech blog with the fourth in a series on best practices in business automation application development.

In the first episode, I talked about the automation imperative that’s driving all organizations to automate. There’s a lot of new technology that lets us automate things that were never possible before, and you need to be considering it or risk losing your competitive edge. In the second episode, I examined best practices in strategic vision including making business automation a strategic direction and picking the right processes to automate that will have the maximum impact. Then in the last episode, I looked at best practices in the design of business automation starting with metrics. Metrics need to be built into automation from the start and need to be aligned with corporate goals. I also talked about how design needs to be based on an understanding of the current process, not a micro level as is analysis but an understanding of the key elements of success or failure in that process. Go back and review those earlier episodes if you haven’t already watched them.

Now in this video, I’m continuing with more best practices in the design of business automation. I want to look at the two sides of automation design that may seem to be in conflict with each other, but they’re actually the most important balancing act that you’ll have during design.

The first of these is: automate whatever possible. If there’s a technology that you can apply to make your processes more automated, you should be considering it. The second is: don’t automate out people where they add value. In other words, don’t get so caught up in the thrill of automation that you remove some essential human contributions.

Just like Goldilocks however we need to know how to apply not too little or not too much automation but just the right amount. So, let’s look a little bit more closely at this.

Starting with the idea that we want to automate everything that’s possible in order to automate effectively. This is driven by some of the reasons that we turn to automation in the first place: efficiency, speed, cycle time, quality. These can all be significantly improved with appropriate automation. Notice I say appropriate automation. And improving these, leads not only to an improved bottom line in terms of costs, but also an improved top line revenue because of customer satisfaction. This is pretty classic business automation design where you look at what you’re doing now and how to make it faster, better and cheaper. That makes things more efficient, it makes them more cost effective and often leads to some improved customer satisfaction for fairly straightforward applications.

Now, automating repetitive tasks is the easiest way to start. So, you look at those individual steps, if it’s done the same time or the same way every time, then there’s almost always some way to automate it. It might be using decision management, business process management or robotic process automation, but the basic idea is to replace a repetitive human step with a repetitive automated step. Then you can look at the more complex automation of less repetitive tasks by using things like machine learning and artificial intelligence. ML and AI can make more nuanced decisions that would normally have to be done by a person who analyzes not only the current transaction, but also understands how similar transactions and decisions were handled previously. In other words, as the automation systems get smart enough to learn from context and past activities, they gain knowledge and skills in the same way that a person does when they’re learning how to do the same task.

Now, you also want to look above the task level. So, you’re not just looking at how can we make the individual tasks automated or not automated. Any sort of broad automation should be looking at the organizational goals and then think about whether things are really being done in the way that they should be. You don’t necessarily have to do it in the way you did in the past. Understanding what you want to get out of a particular end-to-end process, lets you step back and consider alternative processes and methods that may look completely different from what you’re doing.

Now, remember old school process re-engineering where the idea was to radically redesign processes? This is a little bit like that, but with tools that can actually enable a much smarter automated process. So, think about, let’s take insurance claims as an example, think about redesigning an insurance claims so that it has fully automated data and document capture at the front end, and auto adjudication for some simpler claims adjudication. This turns what was previously a very manual claims process on its head, and through the application of a variety of technologies you can have some claims that are completely hands off from beginning to end. Nobody inside your organization has to touch them, they process faster, they have fewer data entry errors, they apply data and decisions consistently… All of this makes the customers happier because you’re getting your job done faster and more effectively and more consistently. They also eliminate a lot of manual internal tasks and this makes your operations more efficient and less expensive. This also frees up your skilled knowledge workers for handling the tough problems and customer interactions that can’t be replaced by automation.

But animation is not a Panacea. You have to understand enough about your processes and your customers to know where your knowledge workers are adding value to the process. Sometimes that’s customer interaction where they’re dealing with customers directly in order to resolve complex problems. This most often happens when something goes wrong with the normal process such as a defective product or a customer profile that doesn’t match the usual pattern, in that case you want to get a knowledge worker involved as soon as you detect that the normal process isn’t working the way it should. Now, you can wait for your customer to type or say “agent, agent” on your IVR, your chat system or you can be proactive and recognize before they even realize it that something needs to be escalated to a person for resolution.

The other main situation for involving knowledge workers is when you have a complex decision that just can’t be made using automation technologies with any degree of certainty or can’t be made yet with automation technologies. If we think back to the insurance example, there are a lot of claims that are just too complex to adjudicate automatically. These are best handled by a combination of automated tasks and knowledge workers.

That last point really brings home the message about design that I’m making in this entire segment. A lot of processes are not all automation are all manual, they’re a combination of both. As a designer you need, to understand what can be automated and what is best left in the hands of the people in the process. As Technologies get smarter, though some of the things that are best done by knowledge workers today will be able to be automated effectively. If you think about auto adjudication insurance claims, this was never possible in the past and now, it’s starting to become possible for more and more complex claims.

We’ve seen this happen in a number of different ways in just the past few years, you look at manual decisions that the what-were manual decisions are now being handled automatically with ML and AI, it really drives home that automation design is a constantly moving target. And you need to understand how new technologies impact your business operations. That means you need to keep up on the automation technologies and see how they might be added into your business operations at any point in time to help improve things without impacting your customer satisfaction. This goes back to what I talked about in the last episode, where you need to understand the key points in your business operations, so that you know what makes them successful and what can make them fail and then consider where the people part of the processes contribute to that, versus what can be automated.

That’s all for today. Next time, I’m going to finish up the design, I knew I was going to do it this time, but I’m going to cover some of the main design anti-patterns or things to look for that indicate that you may have failed in your design somewhere. And then, after that I’m going to finish up the whole series with some best practices in implementation methodologies for business automation.

You can find more of my writing and videos on the Trisotech blog or on my own blog at See you next time.

Blog Articles

Sandy Kemsley

View all

All Blog Articles

Read our experts’ll blog

View all

Learn how it works

Request Demo

Confirm your budget

Request Pricing

Discuss your project

Request Meeting