8 ways to spot the best things to automate
Are you and your team struggling to find the right business processes to automate? Have you experienced automation projects taking a bit too long to deliver or automations that you’ve are launched keep breaking?
You may feel a little nervous that these delays and hiccups could stall the momentum of your automation program, as stakeholders understandably are becoming frustrated and some may have even lost interest, putting automation as a lower priority, as previous excitement has been extinguished.
Perhaps the problem isn’t the technology or even how it’s been developed, maybe you’ve just chosen the wrong processes to automate, and only now are realising how unsuitable these ‘opportunities’ are.
Let me give you eight tips for identifying activities, tasks and processes, which are most suitable for automation. Tips that your teams can start using tomorrow, which can help you accelerate your automation projects and programs, navigating many ‘newbie’ mistakes.
I’ll walk you through some examples so you can see why these tips are so important (and what happens if you miss just one!). once you get a handle on identifying the right processes to automate.
From working in many industries during my automation journey (mentoring and leading analysts and developer teams), building a global network in Asia, North America, and Europe, I’ve noticed patterns in mistakes made, challenges and pitfalls experienced by the many teams I’ve engaged with, and I’ve created methodologies and approaches to navigate these common pitfalls and mistakes that new automation teams suffer from time and time again.
I’m passionate about helping businesses unlock massive savings and thousands of people hours, because of the potential to transform the working-life of staff, empowering them to do more, removing mundane, menial tasks to be able to focus on more creative, problem-solving and human-to-human interaction tasks can indirectly give people a better work-life balance. It has the potential to positively impact both the individual and the business as a whole, and also the customer/client.
I first heard about Robotic Process Automation (RPA) when I was a Process Improvement Analyst (using lean six sigma to improve processes), and back then not many people knew what RPA was/ I found it challenging to find opportunities but through networking, trial and error and reading everything I could get my hands on I create fool-proof methodologies to scale-out automation. But from everything I’ve read, these eight simple tips were valuable throughout my career in identifying where to start:
#1 Standardized/Structured inputs
RPA (“Bots”) can only handle standardized structured inputs, which predetermined like options in a drop down menu or radio buttons. It can’t handle unstandardized data like instructions inside an email message or a comments box.
The reason they need structured, predetermined options is that with free text someone who wanted to write the date they could put 10–11–21. This could mean the 10th of November if you’re in Europe, or it could mean 11th of October if you’re in the States. Even we can get confused as humans, so RPA needs data to be structured.
Gartner had estimated that RPA on its own would be able to eliminate 20% of repetitive tasks. That’s still probably hundreds of processes that you can automate right now using just robotic process automation, but anything else would require some additional AI capabilities.
#2 Doesn’t require human intuition
If a process requires some form of judgment, subjectivity, or a need to understand context this is not a good candidate for automation. Again, RPA doesn’t think, it just repeats a series of pre-defined steps similar to an Excel macro. Machine learning can make decisions using probability, but it’s not a substitute for human judgement, and complex machine learning is not the best start for your automation projects, moreover, machine learning requires lots and lots of data to identify the patterns data accurately.
#3 Rules-based decisions
Similar to the last point, Bots can’t think they can only handle rules-based decisions where outcomes are predefined and the logic to make a decision is programmed. Just as if you were explaining to a brand new apprentice in your organization that doesn’t understand your domain, you will need to tell them exactly how to make a decision.
For example: IF price goes up press “BUY”
ELSE IF price goes down press “SELL”
One caveat here, even if you could determine the criteria for each option, if a decision point had hundreds of different potential outcomes then maybe that’s not the right type of process for robotic process automation. The worst thing about this situation is if you try to automate some of the outcomes, nine times out of ten your robot would probably through up an exception requiring a person to handle the fallout work anyway.
A way around this, if you have a process that is great for automation apart from one complex decision step, you could have a “human in the loop”, where the first part is RPA, then it notifies a person to do some thinking and decision making, and then the person triggers the automation to complete the process. over again and completes that process.
By mapping the process you can get your thoughts on paper, how decisions are made and if they can be pre-defined or not.
These first 3 tips are the bare minimum characteristics of an automation-suitable process, the remaining tips are to ensure you get value from your automation build.
#4 Repetitive & Manually intensive
You want to automate processes that are both repeatable and manually intensive. Look around for processes that have teams doing the same tasks over and over, each iteration taking a while to complete and demanding a significant amount of staff time.
Yes, you can automate processes that aren’t that repeatable and run only once a year or once a month, but it doesn’t necessarily mean that you should automate it. It’s all about the value that you can give back to the business, and if the return on investment is worthwhile. Will the time and costs saved outweigh the time costs spent in building this automation?
Repetitive processes are great for automation not just because of time savings but because our human brains aren’t best suited for doing repetitive tasks, so these processes have a high human error rate. Our brains are best suited to carry out more creative judgmental and problem-solving tasks.
So look for processes that require moving data from one system to the next, copying and pasting, filling out forms or checking long lists of data multiple times a day.
#5 High Volume
Similar to repetitive processes, processes with a high volume of work identify the potential for massive savings. Look for processes where there are long queues, long hold times, low staff morale, teams working late, poor client/customer satisfaction. (This could also be down to poor accuracy of service, but this is also something automation can help with).
If a process can’t be automated directly, like handling customer inquiries, getting customer data to Customer Service Advisors faster, can greatly speed up how they assist customers, as a lot of time is generally spent researching and looking for the right information. Removing this manual work would mean staff can invest more time serving the customer.
#6 A stable process
This tip doesn’t impact your ability to automate the process but it does impact how long your automation would be useful. If a process keeps changing or being tweaked every single day, and if the applications that the process uses keeps changing in layout and structure, then your automated process will not work for long before you need to raise Change Requests to fix things or re-build it entirely.
Robotic Process Automation uses the User Interface just you or me as if it were using a mouse and a keyboard, clicking on buttons, links, writing text in text fields etc. If a button was to move or be renamed, the Bot would create exceptions continuously, and if the process changed and the Bot wasn’t updated it could continue to do something the old (wrong) way.
When designing the automated solution your analyst will need to confirm that the process is always be done the same way, that everyone who does the process does it the same and there’s no chance of it changing anytime soon. Many times I’ve found that team members each have their way of doing something, but with automation, you will need to confirm with the team what the best, most efficient way of doing this process is and standardize the process.
This is a good segue to mention the importance of optimizing the process first before you automate. I’ve always seen myself as a Lean Thinking process analyst first before I’m an Automation analyst. Making a process lean and streamlined before it’s automated not only removes the inherent problems of a process, but it can allow you to get more out of this technology, as you require less Bot-power to automate, and it’s easier to maintain and support.
#7 No pending changes
Adding on to what we mentioned about not automating processes that constantly change, you also want to make sure that a seemingly stable process doesn’t have any pending changes on the underlying applications. Once or twice, I’ve seen huge teams of analysts identifying and mapping out opportunities for automation on long list of processes, spending hours building a huge backlog of processes for their program, all that used a departments core application, paying little attention to the question being asked repeatedly
“Are there any pending changes on this application?”
Only to their surprise, IT notifies them that the pending changes to the core application will put on hold 90% of their automation projects for 6–9 months. True story
To ensure that you can deliver your projects and claim sufficient benefit before any changes are needed you want to ensure there are no pending changes that will start at least three to six months in the future. You also want to make sure that your automation team is connected up to the applications owners so that you’re notified when these changes will happen. This will also allow you to make changes quickly before they’re released, so that you’re not surprised one day when all your Bots stop working.
Yes, perhaps some of your automations use third-party applications or websites and this is out of your control, so in this danger zone if you’ve decided to stomach the risk you can mitigate this by staying in the know, building relationships with the 3rd-party’s tech team so you’re notified of release dates. It’s just best to stay clear of any processes that have applications that you don’t have control of if you can help it otherwise, you’re just going to pay big in support costs and change requests.
The more secure alternative (well arguably the first option above RPA) is to use APIs but they are expensive to build if they don’t exist.
#8 Digital end-to-end
An obvious one, but there are great use cases for incorporating physical robots and the internet of things, which I’ve been discussing with a friend who specialises in Robotics. However, for now, stick to processes that are 100% computer-based, otherwise “human in the loop “ is the easiest option for these use cases.
I hope you found these eight rules useful. Please feel free to leave a comment on these eight if you disagree, or perhaps you have more/other tips for identifying good automation use cases. I’d love to hear your comments and feedback
If you want to learn more about Intelligent Automation in your office, subscribe to my YouTube channel Tony IA (Intelligent Automation, Simplified) for videos created weekly, to simplify intelligent automation for business leaders and professionals who are new to automation to level-up your knowledge. Become empowered so you can start optimising your business and discovering new technologies, in a lean and accelerate way. You can also learn more from my book, Business @ the Speed of Bots: The AEIO YOU method HOW TO IMPLEMENT ROBOTIC PROCESS AUTOMATION THAT SCALES. Get ready for the new digital transformation age for more information. The foreword is written by Guy Kirkwood, who is the ex-chief evangelist at UiPath, and a very well-known advocate of RPA with over 20 years of experience in outsourcing.
🔥🔥 [FREE EBOOK] SIGNUP TO THE NEWSLETTER TO GET A COPY 🔥🔥
#1 strategy for accelerating your career in the world of Process Automation and Intelligent Automation to become an expert your team and business can turn to for guidance