The Secret to Achieving Quality Output from Robots

Posted: 16th April 2018

Higher quality output is one of the key drivers behind automating processes using robotics, but how can firms ensure they realise the benefits?

The process a robot is given to run is always adhered to, and there will be no variance in how it fulfils its task, unlike within manual processes, which are naturally subject to some human error. When processes are completed manually, this natural degree of human error or omission is often not identified, or only identified when the error has a negative effect on customer outcomes.

The complete adherence to process achieved by robots leads to more reliable outputs from processes. However, if the processes and policies are not fit for purpose in the first place, robotics and workflow tools will never be able to make up for this shortfall or enhance the quality of the output above the maximum level the rules allow.

What to consider before automation

When automating a process, the first step has to be robust internal challenge; scrutiny of the process to determine whether it really is set up to achieve the firm’s desired outcome. Automation provides greater opportunity to challenge and reengineer processes – more complex options become operationally feasible when not constrained by the cost of the human resources needed to undertake them.

At the start of an automated workflow project, it’s important to ask the following questions:

1. At 100% adherence, does the process give enough confidence in the output?
2. Have process enhancements been discounted historically due to manual resource constraints?
3. Is the process currently compromised due to cost or time?
4. What is the ideal outcome?
5. What options are there to move closer to the ideal?

Once the answers have been established, it’s possible to reengineer a process to provide higher quality output using robotics and workflow management, whilst delivering the cost and time efficiencies that primarily drive the need for automation.

An example – Know Your Customer (KYC) screening

An organisation has been struggling with its KYC screening for new customers. Their current process involves manually checking four data sources, including Google, and creating a report in ‘Team One’ to be assessed by their more experienced colleagues in ‘Team Two’. It takes one person in Team One over a day to carry out the research and compile the reports for five customers. Any errors identified by Team Two are then fed back, and the process is rerun.

The answers to the first four key questions above are:

1. There are not enough data sources being checked to give a thorough view of the customer. Many investigations are unable to give either a clear pass or fail, meaning that a second review and additional data checks are carried out on approximately 40% of the cases.
2. Yes, the four basic checks were chosen as they can be completed in a reasonable amount of time. Adding additional checks would extend the time it takes to carry out the activity on 100% of cases, rather than the 40% it is carried out on at the moment.
3. Yes, the impact on the time to approve applications and the additional resource costs have driven the decisions on the current process.
4. All cases would receive the more robust additional data checks – with 11 data sources – and there would be no reruns due to poor quality being identified, alongside a reduction in the time to complete.
    

And on the fifth question, what options are there to move closer to the ideal?

In this case, robotics opens up the opportunity to achieve the firm’s ‘ideal solution’:

• The 11 checks can be carried out in parallel rather than in a linear process, significantly reducing the time required for the checks to be completed – additional checks can be added without increasing the time required
• The reports will all be 100% reliable (as per the rules set out for them) as information is transferred without any loss of integrity due to human error
• Reruns are removed and the number of cases requiring additional work will be dramatically reduced
• The data sources can be changed, added or removed with no training requirement. As soon as the new policy and process is rolled out to the robots they will immediately start following it

Without robotics, the process would continue to be constrained by time and resource. Using this process, Team One can be reallocated to other tasks or upskilled to support Team Two in carrying out the subjective reviews of cases to increase the volume processed each day.

Using this new option would enable improved outcomes – both in terms of accurately identifying potential bad actors and preventing them becoming customers, as well as providing a quicker customer journey for customers, with fewer false positives causing delay or upset.

Robots are only as good as the processes they are given

The key here is that automating a process which is not optimal will only ever drive the same outputs (albeit more quickly and consistently). To achieve the real benefits from a robotic automated workflow process, it’s important to look more widely at the opportunities the time and cost efficiencies of automation provide.

Where possible, ‘compromised’ processes (having previously been limited by human resource requirements) should be reengineered and enhanced to provide not only cost and time benefits, but genuine quality improvements.

Making automation of processes truly successful requires a mindset shift. Fortunately, however, this is a mindset shift which removes many of the traditional limitations of process optimisation, and allows businesses to get far closer than ever before to ideal customer outcomes, in a more consistent and more cost-effective manner.