Skip to main content

Scaling intelligent process automation – how to do it better

(Image credit: Image Credit: MNBB Studio / Shutterstock)

This pandemic is generating unprecedented operational demands on global organisations, so the need for Intelligent Process Automation (IPA) technology and its ability to drive transformation faster - with fewer resources than before – will only be greater as we hunker down, then begin to emerge.

One of the biggest problems for organisations when automating their business processes with IPA is how to sustain and scale up these activities. The reality is that not enough business automation projects are really progressing past infancy and merely getting the technology in house and grinding out a couple of bots does not really move the needle. You need a wholistic IPA practice that delivers consistent, repeatable results on an ongoing basis.

The people, process and structure behind IPA are more important than the technology itself. The key here is having a robust delivery methodology right from the start. These are the critical factors that determine success and failure. Below are several steps that must be undertaken to ensure that your IPA practice becomes an ongoing value generator across your operations.

Align with a vision 

For any IPA practice to have a chance to succeed, we need an explicit strategy and purpose. An IPA practice has tons of benefits, such as improving data quality, operational efficiency, process quality and employee empowerment, or enhancing stakeholder experiences by providing quicker, more accurate responses. Save cost, improve quality, blah, blah blah…  All of these are important, but they none of them are exactly flags to rally around. The vision of your IPA strategy should be aligned to wider needs of the business and should be at least somewhat inspiring. Here is one idea to get you thinking:

We elevate our personnel by enabling the business to leverage our humans for what they can do uniquely well, leaving rote, mundane, repetitive tasks to the machines. We do this by leveraging RPA and IPA technologies in a reliable, repeatable way to deliver measurable and meaningful results to the business.

It is critical to secure the backing of key stakeholders. When you encounter a roadblock getting robot IDs because HR says that they need a headshot for the parking pass they are going to issue (yes, this is a real example),  you need someone in your corner with the authority to change that process, because it hardly applies to robotic workers.   If your stakeholders see RPA as a strategic business project, they will support it – and provide the necessary financial and human resources too. IPA is always a joint endeavour between IT and the business, so both parties must be involved at the beginning.

IT will support IPA on many critical fronts, such as compliance with IT security, auditability, the supporting infrastructure, its configuration, and scalability. The business provides the backlog of Opportunities, and the business framework necessary to close the value loop on the automations.

Get organised

There is a great deal of handwringing in our industry about where and effective IPA practice needs to sit within the business. Should we use a centralised approach? Federated? Does it sit under IT? The business?

There are two critical things that we must get right here, and everything else tends to work itself out. The first of these is that we need to pick an approach and then execute. If you get the fundamentals right, any approach can work, you just need to pick one and go.

The second, and more important thing is to stop thinking of a centralised vs. federated vs. siloed as competitive models. They are not. They are different flavours of the same approach, at different levels of maturity.

Most of the very successful practices that I have been involved with have started with a centralised model. You cannot brute force your way into a successful IPA practice the way you can with, say, a new version of SAP. As I mentioned before, cultural adoption is even more important than the technology that you are putting into place. A centralised model allows you to build the best practices within the organisation in a way that can scale later. And it lets you stumble while you are still small.

This means that you need to have an IPA Centre of Excellence, but it does not matter where that sits, just that it owns the business result. I have seen successful practices live in IT, in specific lines of business, in shared services organisations, or even outside the company altogether.

As you mature, you will start to federate responsibility for portions of the automation practice. We like to think of a successful automation practice as consisting of three main activities: Opportunity Identification, Development, and Operations (more on that in a future post). The easiest of these to federate is Opportunity Identification. As you mature, you will find that you can federate more of Operations, and even Development to your lines of business.

Indeed, the “Citizen Development” model that there is so much noise about these days is simply a vision of extreme federation, where even the development process is given over to individuals within the business. This can work, but it is extremely tricky to get right, and you better have the appropriate level of maturity before you make the attempt.

Many businesses make the mistake of starting out federated. This is a bit like learning to ride a bike without training wheels. It can absolutely work, but it is quite a bit more effort, and you will cause yourself a lot of unnecessary pain. It is a common mistake, and it stems from thinking of centralised and federated as two distinct, incompatible models. Stick with the maturity concept, and you will have better outcomes.

Avoid lapsing into a divisional approach at all costs - where multiple RPA functions run separately across the organisation with differing infrastructure, governance, and delivery teams. Naturally, this ‘siloed’ non-collaborative, poorly controlled set up is guaranteed to prevent you from achieving true scale. This is basically the same thing as having several Centres of Excellence, all of which vie for dominance. It is wasteful, and while it is possible that you could end up with one of those centres gaining a foothold, it is far more likely that they starve each other of the resources they need to survive, and none of them thrive.

Selecting the best processes

Are you struggling with your IPA journey? Consider having us do a maturity assessment to help tweak your process and provide actionable insight to align with the latest industry best practices. Read more here. Need to upskill your crew? Look at some of our training offerings here. Finally, we are expanding! Do you strive to be among the best in the world at delivering IPA value to businesses? Consider joining our team.

When identifying process automation opportunities – always select ones that will generate the fastest benefits and be crystal clear about what makes a truly good process. Even in the very beginning, you should have a good idea of the benefits and potential cost of the opportunity.

The best options are those processes involving a high volume of manual and repetitive tasks, or suffer from human errors, or require customer experience improvements – such as faster response times. And for goodness sake, please, please, please, select easily built processes for your initial several opportunities in any line of business. The biggest cause of failure I have seen is initial selection of large-benefit processes that take months to develop. You need to get something into production in a few weeks. Prioritise thusly.

Prior to automation, think carefully about streamlining or improving processes as this requires additional time, cost and effort and usually results in only minor improvements to the automated solution. Any decision should always depend on the overall strategic business objectives for automation. While it is true that it is best not to automate a bad process, it is equally true that it is better to get something automated than to sit around pontificating about how to lean out the process. Again, get something to production fast.

Next up, talk with the IT department about any maintenance planned for the target applications, and if process automation should wait until that is done. Finally, make sure that you will capture the benefits. Agree on a set of indicators; financial, process, quality, and performance related KPIs.

Consider how to generate demand for IPA within the business. Proven routes include defining automation champions, running internal communication programs and workshops for engagements, and providing employee incentives for identifying suitable processes. Also, establish meaningful measures of IPA’s value such as how many hours saved - and what they accomplished and alignment with core strategic business metrics like contribution to operational efficiency and employee retention.

Deliver with care 

To start automating processes, capture the correct information in the define phase to avoid problems, so involve knowledgeable subject matter experts in this activity. It is also worth holding a process walk-through for the right audience. Each chosen automated process must be documented and an understanding gained of how it will differ from the same human process. Once this has all been agreed with the business, and the process design authority has approved the proposed blueprint and conducted the necessary peer reviews - development can begin. Only when the business is satisfied, can sign-off testing begin.

Maintaining operations 

Once processes are in production, they need the right support around them. Ensure that Digital Workers are handing back business referrals, or exceptions, to the operational team for manual intervention, and that a technical capability is readily available in case the Digital Workers do not act as expected. Ultimately, to ensure continuity and availability of automation resources, there must be a robust, supporting IT infrastructure. Give some thought to how you will monitor for the end-to-end result you are striving to achieve. (More on this in a future post)

Train and retain the RPA teamx

Appointing high-quality lead developers is essential to deal with the inevitable IPA pain points but these people can be extremely difficult to find in the current market. Developers will need to be trained to the highest standard in development and process analysis. This way, they can perform in a hybrid role, while receiving on-site support from an experienced consultant. As a development team continues to grow, appoint a Design Authority to ensure standards are maintained and a Control Room Monitor to manage the production robots.

Having the best talent in IPA is the life blood of any successful initiative and with these skills being increasingly more sought after has led to demand outweighing supply. Therefore, it is worth considering a partner that provides the human resources, governance, management, and methodologies to support global learning programs and integrate training into rapidly expanding IPA implementations.

In our engagements, we will source, train, and mentor a fully accredited IPA team of people within weeks through our ScaleSafe™ model. This team is then committed for several years and are equipped to maintain a dynamic, sustainable centre of intelligent automation excellence.

Final thoughts  

Ultimately, to gain the best results with IPA, we must define the complete journey upfront, rather than waiting for mistakes and then correcting after. Once we gain company-wide support and create a vision of desired results, start delivering, and fail fast. Allow the IPA program to really thrive as it scales across the business. Get this mix right and you will create the platform to explore and build more intelligent automation offerings - using AI and other emerging technologies – to maintain innovation and sustainable success. You will be climbing away, while most other people will be struggling to stay aloft.

Shaun Dawson, CEO, Robiquity