You are here


Automation goes beyond tech to transform your business

A company needs clearly defined objectives before deciding on the best tech that suits its needs.

Given that automation technologies are relatively new fields to a typical business stakeholder, CFOs will likely need to invest more of their time in stakeholder engagement and working collaboratively across the business, particularly with IT.

FOR years, we have been saying that automation is coming. It is firmly here today, yet gaps in understanding of automation technologies continue to exist among business leaders - and not surprisingly, given the pace of technological change.

For some, automation may immediately conjure a limited view of physical machinery or a grandiose vision of robots taking over the world. The danger with such views is that it almost always makes automation a technology project. On the contrary, automation should first and foremost be a business-driven exercise.

Instead of hoping for or adopting a one-size-fits-all approach and solution, organisations need to specifically define their business objectives, select the processes to change accordingly, before finally deciding on the relevant technologies to help achieve those objectives.

The possibilities for automation differ greatly between functions: The finance function has the highest potential given that as much as 80 per cent of its tasks could be automated, according to the EY report The future workplace: How to automate intelligently.

Faced with a plethora of automation possibilities, where should CFOs begin?

Your feedback is important to us

Tell us what you think. Email us at

Prioritising automation starts with increasing their understanding of the different ways that automation could impact the business, ranging from the more known outcomes of greater labour efficiency and increased speeds, to lesser explored advantages in capability expansion and quality improvement.

Don't confuse RPA with AI

To automate standardised, simple and rule-based activities, many organisations use robotic process automation (RPA). RPA is widely used in finance automation projects, such as basic report generation, and bank and inter-company reconciliations. In particular, finance processes that are high in volume, repetitive and rule-based, such as accounts payable processing and accounts receivable processing, have great automation return on investment.

However, as every activity automated using RPA needs to be individually coded by a developer, using RPA to automate processes that are complex or involve unstructured data is time-consuming. In these situations, organisations look to cognitive automation, which is a combination of RPA with a field of artificial intelligence (AI) known as machine learning.

At this point, it is worthwhile drawing a distinction between RPA and AI for clarity: the former is a technology while the latter is a field of computer science.

AI project set-ups are completely different. It would involve either setting up a portal so that the machine can "watch" what human operators do, or would require the function to feed the machine with lots of historical examples that allow the machine to create its own code using an internal algorithm.

In our work with clients, we have seen a project to automate an end-to-end order management process using machine learning. A robot "learnt" to recognise several different types of unstructured documents such as invoices, bills of lading, import and export notes. It then further learnt how to extract 70 different types of information and enter these into several different systems while executing a number of critical controls.

The ability to perform more complex tasks may give the impression that AI is superior to, and could even replace, RPA. However, AI's capabilities must be weighed against its demands on implementation resource and time. A simpler and quicker solution like RPA could well work for the many rule-based tasks that exist in any function.

Regardless, there are still many under-utilised potential of RPA and AI in the finance functions. This includes areas such as financial planning and analysis, where they could be deployed to automate the pre-population of forecasts using historical and market data, loading pre-populated balances into the planning system and creating variance reports.

For cognitive automation, technological advances will drive greater sophistication in usage and tasks that RPA alone cannot enable, such as using AI applications to identify fraud and irregularities in payment processing, or to improve risk management and control.

Don't forget organisational transformation

Each of the automation use-cases described will reshape processes, jobs and the function's structure. As with any technology transformation exercise, businesses will need to deal with implementation issues spanning stakeholder management, change management, resourcing, project scoping and timeline management.

However, given that automation technologies are relatively new fields to a typical business stakeholder, CFOs will likely need to invest more of their time in stakeholder engagement and working collaboratively across the business, particularly with IT.

Selecting the right automation project becomes even more critical for smaller companies with limited resources. As AI projects generally require a larger investment, consider RPA projects first that are less costly and can be rolled out in a matter of weeks, which can help to establish a use-case.

One way to garner confidence would be automating a simple process as a quick win. Processes that are highly standardised, voluminous and repetitive are particularly suited for automation use-cases.

CFOs will need the support from within their teams as well. Finance teams will need to be open-minded and change-ready. Having the right mindset will motivate team members to set aside time and effort to contribute to the change process, such as in defining business rules and exceptions handling that are essential input for the automation project.

Translating such information into a reader-friendly format for the technology will require the expertise of data scientists and software developers. For smaller companies that are interested in exploring automation, the lack of access to such in-house technical capabilities is a dominant barrier.

This can be overcome by providing training in coding skills or exploring partnerships with other like-minded organisations to offset the cost burden or accelerate project implementation. These collaborations to co-develop or co-implement automation technologies may also be eligible for government grants and funding from various industry bodies.

RPA and AI are indeed some of the most exciting technologies today. As the number of automation projects increases, business leaders and CFOs will do well to self-check if they are merely automating blindly or in silos, without critically considering their strategic business objectives, the suitability of the technology, and the organisational transformation the project entails.

To succeed in their automation journey, organisations must think beyond technology and recognise that automation is primarily a business transformation opportunity.

  • The author is head of advisory from Ernst & Young Advisory Pte Ltd.
  • The views in this article are those of the author and do not necessarily reflect the views of the global EY organisation or its member firms.


BT is now on Telegram!

For daily updates on weekdays and specially selected content for the weekend. Subscribe to