Finding the right tasks to automate can be a challenge. There will be business pressure to target the high value inefficient activities, and a perception that the path to automation will be a seamless transition all handled without any disruption to business-as-usual activities. The reality may not be as rosy and so it is important that the most appropriate tasks to automate are chosen to give the best chance of success.
When we look at tasks to automate, we often evaluate them using the following criteria:
- Is the process repeated?
- Is the process stable?
- Is the process structured?
- Are there less than 20 steps to the process?
- Is the process well understood and rule-based?
- Can the process complete without subjective decision making
A task that isn’t repeated, stable or structured is unlikely to ever be a good candidate for automation. Whilst the toolsets do provide a level of flexibility to accommodate small changes if the core task is always changing the risk of operational errors and downtime is likely to be high.
After identifying a process that meets the above criteria, the next question is does the business still need to run it. The most successful saving of time is finding a process that you can stop running. So many tasks are completed by teams simply because that is what they have always done.
When choosing a task to automate we should first critically assess whether the task even in its current form is still adding value to the business. Rarely do we look critically at the activities that we do as part of business-as-usual functions and ask if they are still needed, and if they are needed is this the most efficient way of achieving them.
Evaluate, refactor, re-evaluate
Taking a manual task and automating the steps exactly as performed by a user is unlikely to be the most optimal solution. Once we have agreed a process is still needed or is valuable, we review the steps and evaluate if the manual process could be improved. Why automate a 12-step task when with some simple refactoring we may only need to automate a 5-step task.
After refactoring the task and stripping it down to its most efficient version of a manual task we should ask ourselves again is it still worth automating the task? If it is we have a simpler, more efficient task to automate. If we don’t, we have reduced the complexity of the manual task and saved ourselves the development cost of automating.
This isn’t a failure of automation, quite the opposite, we have used our automation mindset to save the business time and money with the least amount of effort and resource draw.
So which tasks go first?
It can be tempting to go after the big prizes first. To showcase your new automation function by tackling the biggest problems and come out of the gates storming. If it was this easy though there wouldn’t be any big issues left. It is better to start small. Build up your team’s experience by focusing on small achievable automations within a supportive framework.
As your skills and confidence grows so too can the automations the team can deliver. This methodology of creating small automations is also a good approach to automate a large complex business function. When solutions are tightly coupled with huge, interconnected processes the risk of operational failure is high. A larger automation that carries out all the task within one solution is likely to have a high risk associated with it. Creating an end-to-end solution using smaller tasks all orchestrated together is more likely to succeed.
What do you mean by smaller tasks?
If we look back at our original 12-step process mentioned above, which through refactoring we have now redesigned to a 5-step manual process of:
- Check to see if email from accounts department has arrived.
- Download email from account department and place in correct directory.
- Download billing information.
- Cross check accounts fill against billing department information.
- Submit reconciled information to financial portal
We could automate this into a single 5-step automation process. Or we could subdivide this process and separate the submission to an external portal into its own module. By doing this we reduce the risk by reducing the complexity of the automation and we can reuse these small modules across multiple workflows. So instead of creating 20 workflows that all have their own portal submission sections we have 20 workflows that all call a common portal submission task.
Are there tasks to avoid?
One of the areas to tread very carefully with are those that staff members either enjoy doing or are tied to their view of their worth to the company. When staff members perceive their value as aligned to a single function or area they can become very concerned at the prospect of some or all of that function being transitioned to an automated task.
There will be times where this is inevitable because the manual task is causing a bottleneck or is too expensive to run. However, when these scenarios emerge it is important to have a clear vision of how the individual’s workload will adapt post automation and ensure that they see a role that is of at least equal value to them as the role prior to automation.
Can we just get someone to tell us what needs to be automated?
There are many companies who can help support organisations to identify and assess the business processes that might be suitable for automation. These outside organisations can bring a wealth of automation knowledge that can support decision making. But when picking a partner it is important to pick one who has experience in your business area, and ideally one that you have a long-standing relationship with.
The rationale being how much can a third-party organisation really understand about your business and its challenges in a matter of weeks or a couple of months, if they have never operated in your environment? By partnering with organisations with deep subject matter expertise in your business functions and operating model you are not paying for them to develop knowledge about you, and instead they can focus on the brief that you have set.
This doesn’t mean that you can completely outsource this function, but if you need support to help assess the potential for automation there are many options available.
Should we just get someone in to do the automations for us?
The barriers to creating and using automation tools and technology have significantly lowered over the last few years. The creation of content is no longer reserved for specialist coders but has been opened up via the use of drop and drag features and prebuild system components. That being said it is important to have individuals with a development background and an understanding of the technology to enable high value automations to be delivered.
As an organisation you will need to decide whether to invest in your teams to grow this technical capability within your service lines. Or alternately reach out to the wider market to find partners to help deliver the technical components you need. There is no right or wrong answer, but it is important to understand the choice between growing capability in your own teams and buying specific pieces of work.
How can SCW support us?
Being a technical leader in the Health and Social Care landscape SCW has significant subject matter expertise and technical capabilities. We can provide expert advice on automation tools and techniques, and advise on how they might optimise your business processes.
Our teams have been focusing on the development of automation frameworks and tools with the potential to deliver value across health and social care settings.
For more information contact Thomas Counsell, Deputy Director of Research and Development, SCW Innovation,
