Top five challenges in business process automation
Business process automation is now mainstream, gaining acceptance in the marketplace and experiencing rapidly increasing adoption rates. Opportunities for process improvement, improved data integrity, labor cost savings, prompt ROI and improved employee morale are just some of the many reasons for this growth.
The maturity of process automation technology and platforms has enabled this trend and created these opportunities, along with some intriguing, eye-popping success stories.
However, studies by EY and Deloitte in 2019 provided the cautionary tale, i.e. that approximately half of RPA projects fail, and most of those that succeed do not deliver the expected ROI.
Investing thought, time and effort into the following challenges can help ensure that your automation initiatives are successful, achieving your goals and delivering expected ROI.
To succeed, you need to know what success is and how to measure it. We encourage all of our clients to define success in preparation for the Kickoff Meeting.
The definition should include high level goals along with some specific goals at the granular level. We will use AP Invoice Automation as an example of a business process, since it is very common.
First, look at some universal numbers, dollars and percentages for AP Invoice processing:
- 40% of all companies do NOT measure the cost of an AP Invoice
- Cost to process AP Invoices ranges from $ 2.07 to $ 10.00 per invoice
- 17% of all invoices have exceptions (Exception handling discussed later in this article)
- One FTE can process 1,000 invoices per month (includes all components of AP processing)
- 1.6 % average data entry error rate for AP Invoice processing
Second, measure numbers, dollars and percentages for your existing manual AP Invoice system:
- Time spent per month processing (include all FTEs, part timers, temps)
- Cost to process AP Invoices (if you don’t know the formula, reach out to us)
- Estimate the percentage invoices that have exceptions
- Estimate the data entry error rate
Third, compare your numbers, dollars and percentages against the universal numbers, dollars and percentages in order to set your measurable goals for:
- Time spent per month processing AP invoices
- Cost of processing an AP invoice
- Data entry error rate
- Return on investment
The above part of your Definition of Success represents hard costs and benefits. Now consider soft costs and benefits, such as operational efficiency, improved employee morale (automation can increase this by up to 43%), and workforce repurposing.
Conclusion: Visualize the end of your automation project and how to determine if it was a success. Do not minimalize this step. Performing this step will provide accountability and show you areas of improvement for your next automation initiative.
Selecting the right business process to automate is always of utmost importance, but if it is the first business process your team, department or organization is automating, it is even more important.
There are many different factors to consider when selecting a process, and the highest priority factors will change from organization to organization. Some factors may be organization specific, consider:
- Corporate G&A reduction initiatives
- Increasing data entry workloads due to corporate growth
- New business process(es) not currently assigned to human workforce
- Automation considered part of Corporate Digital Transformation initiative
- Plugging a gap in the workforce (Automate the workload of the person who left the company)
Now consider factors specific to automation, i.e. business processes that are well suited for automation regardless of the organization:
- 20+ hours a month
- Clear business rules
- Manually intensive, high volume
- Current labor rate for process is high
- Low exception rates and low error rates
- Application(s) in the process are not subject to frequent change
- Input data is structured (this is ideal, as opposed to semi-structured or unstructured input data)
NOTE: There are tools in the marketplace that walk you through many questions to help determine the suitability of the process you want to automate along with a cost/benefit analysis.
The factors specific to your organization will highlight specific needs of your company, whereas the factors specific to automation will give you a clearer picture of your chances for success.
ROI will often be the key factor in deciding to automate a business process, and it is our belief that each automation project should “stand on it’s on two feet” and deliver a return within twelve months, which current automation technology and functionality can do.
However, there will be processes that may not deliver the highest ROI but may still be chosen because of immediate corporate needs, such as plugging a gap in the workforce because of turnover.
If you consider all of the above, you will be able to build a true business case for each business process you are considering for automation, and set more precise expectations with your team and management.
Conclusion: After considering all relevant factors, you will be positioned well to make a business case for the process you want to automate, providing expected ROI and your chances for success.
The business process you plan to automate is typically being performed manually by front line member(s) of your workforce, referred to as subject matter experts (SME).
These individuals are key to the success of the project, playing a leadership role in explaining how the process works, the business rules, exceptions, frequency, required authorizations and dependent factors. It is these same individuals that will perform user acceptance testing (UAT).
It is ideal but not always possible or practical to select business processes that are currently being performed by individuals that have leadership qualities, a positive can-do attitude, and good communication skills.
Communication between these individuals and the resources actually building the automation is crucial, and can either shorten or lengthen the overall lifecycle of the automation project, depending on how clear, specific, thorough and responsive the communication and listening is.
Some cultural sensitivity from management can go a long way regarding the SMEs, as sometimes they are wary that their jobs are being replaced by automation, and in some cases it is actually true. Handling this sensitive matter with integrity and common sense can be the difference between true collaboration or total resistance and lack of cooperation.
The progressive, agile corporations today evaluate their workforce carefully, and then invest time and effort in the repurposing exercise. Two years ago executives listed repurposing as their biggest challenge regarding automation, and ideally the workforce who now have additional hours available in their business week have a clear direction on how they will spend that time.
Common repurposing functions and activities include increased collaboration with customers, suppliers and team members, analysis, innovation, and retraining for new roles. It is important early in the automation process to communicate to the workforce that the automation of their tasks represents growth and opportunity for them.
Some companies have reduction in headcount as their primary goal, other companies want to maintain status quo with headcount and have automation handle new and increased workload, so be sure to include the impact automation will have on headcount in your project charter and definition of success.
Conclusion: SMEs are crucial to the success of the project. Their leadership, expertise and communication can impact the project in a positive way. Be as clear, transparent and sensitive as possible regarding the impact on headcount, as human beings will be affected. Plan well in advance how the company plans to repurpose the affected individuals, structure it into a program if possible, identify new ways in which your affected workforce can add value to the organization.
The 80/20 rule applies directly to exception handling. Practical decisions and executive decisions must be made carefully regarding how exceptions will be handled. Your decision on handling exceptions will directly impact your project timeline, specifically impacting the time it takes to build the automation and also test it. Here’s how and why…
An automation consists of a workflow with business rules and exception handling. Using the AP Invoice Automation example again, let’s say 80% of 1,000 invoices being processed complies with all business rules, meaning all the required information is there, the supplier is present in your billing system, dates are correct, the part number and price in the line item detail are correct, etc.
This means 800 invoices can be processed automatically. How do we handle the 200 invoices that were not able to process automatically?
One option is to send all 200 invoices straight the AP Manager, and let her or him figure out what to do, how to handle the invoice manually. This option requires the least amount of code when building the automation, and the least amount of testing.
Another option is to build in exception handling for common errors in your invoice processing. Let’s say the Buyer is wrong, the PO number is not on file, or pricing is incorrect, and these three scenarios happen frequently, let’s say half of all your exceptions . You can build exception handling for all three of these scenarios, and send the exception invoice directly to the person best suited to handle that exception, rather than straight to the AP Manager.
This will streamline the process, and enable your AP department to get those invoices corrected and processed quicker. It would also NOT impact the timeline negatively, as the coding of three exception handling routines would not require substantial time.
The lesson here is to build in exception and error handling for scenarios that happen frequently, as that increases the number of transactions that are totally automated, as opposed to coding exception handling for scenarios that might happen on one invoice per month.
Exception and error handling built into the automation is a powerful way to recover gracefully while processing many transactions, in other words continue with the next transaction, and also log details regarding the exception and send email notification(s) to the appropriate resource(s). Use this feature wisely, as it directly impacts how long your project will take, and more importantly, how long the new business process will take.
Conclusion: Theoretically, you can build in exception handling for almost any exception that happens when processing transactions, but is it worth it? Exception handling for errors that happen frequently are worth the effort, creating a more streamlined business process.
Testing the automation of business processes resembles testing of ERP systems. Long ago, there was a survey that stated 50% of ERP projects fail because of testing. The same characteristics that caused the failure rate in ERP projects are alive and well in automation projects.
Because of this, you should approach testing your automations in the same manner as you test your ERP systems, dedicating sufficient time for hands-on testing, creating test scripts, and preparing a non-production test environment with.
Dedicating sufficient time for hands-on testing – We include the partner or the internal tech team at the beginning of UAT, as they may be able to get the SMEs past some minor issues when first starting. At a certain point however, the SMEs should be able to process all scenarios completely on their own. SMEs should also be performing some benchmarking while executing these scenarios, e.g. time of day, how long the automated process ran, and latency issues.
Creating test scripts – Well thought out test scripts will usually include up to 20 scenarios. It’s not always possible to document every scenario, so to be efficient focus on the most common scenarios, in terms of frequency and volume, and also scenarios in the critical path, or that can stop other processes from executing, in order to mitigate risk.
One challenge we commonly see is related to the non-production test environment. In some projects, there simply was no test environment, so we were forced to test in production, not ideal and in violation of best practices, but it can be done if necessary.
In other projects the environment was so much different than the production environment in size and latency that we actually had to change code regarding the launching and logging in to applications due to timeout issues in order to roll the solution out in production.
Along with drawing out the overall project timeline, this in fact meant we were now testing in the production environment.
Note this section purposely focused on the UAT performed by the SMEs, and did not address the unit testing and integration testing performed by your partner or internal tech team building the automation.
Conclusion: In order to succeed, you will end up spending a lot of time testing your automation. The ideal environment and time to do this in is your non-production environment during UAT. If you do not test thoroughly there, you will end up testing in your production environment, which is not ideal and can actually stop you and/or delay you from processing business transactions.