Over the last two decades, there is irrefutable evidence that automation is having a major impact on nearly all technical industries. Stephen Brennan and Ken Hardy of our R&D incentives practice explain.

Automation comes in many forms; from simple point scripts designed to shave seconds off monotonous tasks, to large scale process automation which has the capability to save millions of euros and significantly reduce the resources required to complete a task.

Example 1: Oil and Gas

The oil and gas industry is changing. While their employee headcounts are increasingly shrinking, the number of rigs climbs. Why? The answer is robotisation and automation, which in this case has largely been driven by migration to automated oil rigs, leading to a significant drop in overall employee headcount per rig.

The technology required to execute this automation task is far from trivial. For example, electrical engineers are automating oil rigs by deploying robots to monitor equipment and potential leaks. The design and test of these robots requires significant investment into research and development teams and will likely involve the application of cutting-edge technologies such as machine learning, computer vision or neural networks.

Figure 1: Graph showing declining overall Employee Headcount and increasing overall Rig count (Durden, 2017)

Figure 1: Graph showing declining overall Employee Headcount and increasing overall Rig count (Durden, 2017)

Example 2: Software Development

Let’s also consider a different application of automation within industry – functional testing in software development. If no automation framework was in place, a human test engineer would be required to invest many hours in product testing. This generally involves validation of the mainline functions, basic usability, accessibility, and error conditions.

The benefits of automating software testing are multiple, including improved product, reduced testing time, increased reliability and fault detection, as well as reduced costs and time spent by humans on more menial tasks. It also significantly reduces the resources required to reach each next testing milestone, while also increasing the speed at which the product is tested.

Two examples of automation are described above and yet from the perspective of the R&D tax credit, one of these two cases would be more open to challenge than the other in the event of a Revenue audit, even though both are automation-based projects – why?

Automation and the R&D tax credit: No Silver Bullets

To understand the nuances that may play a part in determining the qualifying or non-qualifying aspects of automation it is important to understand how automation, by definition, is perpetually located in a grey area within the R&D process as dictated by the legislation (Section 766 TCA).

Several questions can be asked which will determine whether your automation activities qualify for tax relief as outlined in the legislation.

  • Is the activity performance in a field of science or technology?
  • Is the activity systematic/investigative or experimental in nature?
  • Does the activity encompass one or more of the following? A. basic research; B. applied research; or C.  experimental development
  • Does the activity seek to achieve technological advancement?
  • Does the activity involve the resolution of scientific or technological uncertainty?

If the answer to any of the above is ‘no’ then you may not qualify.

What is automation?

Automation is defined as a technique, method or system of operating or controlling a process by highly automatic means, reducing human intervention to a minimum.

The element of the definition above that gives rise to confusion about the qualifying nature of automation-based projects is the use of the word “process”. Does this refer to the automation of a known process or the modification and optimisation of a process in which automation plays a key role?

This leads to The Process Question:

“Did the activity carried out involve a change or optimisation in the process that gave rise to technological uncertainty and potential technological advancement?”

It is only by answering this question that we can perhaps gain insight into the potentially non-qualifying and qualifying automation-based activities.

Consider the following: In the table below, there are five questions that must be affirmed in order to satisfy the requirements for the R&D Tax Credit. If we return to the two examples discussed above and run them through this flow chart with emphasis on The Process Question, we can begin to understand why one of the examples might be more likely to be challenged in the event of an audit by Revenue than the other.

Question:  Example 1: Robot to monitor equipment and report potential gas leaks   Example 2: Automation Of a Functional Test Plan 
Is the activity performed in a field of science or technology?  Yes: Electronic Engineering, Computer Science   Yes: Computer Science 

Is the activity systematic/lnvestigative or experimental in nature? 

Yes: Experiments required to validate that the robot can 'see' gas leaks, accurately report when and where leaks occur etc   Yes: Experimental activity is required to ensure that the automated tests are running correctly 
Does the activity encompass one or more of the following:
1. Basic Research
2. Applied Research
3. Experimental Development 
Yes: Experimental Development  Yes: Experimental Development
Does the activity seek to achieve technological advancement?   Yes: This project does not simply involve the execution of a known process by a machine. This project requires a fundamental change in the process in order to be successful

No: This project simply involves the automation of a known process, meaning that no knowledge has been added to the field of computer science 

Does the activity involve the resolution of scientific or technological uncertainty? Yes: In this case it is not feasible to directly automate the process executed by a human (walking to inspection areas, visually inspectlng for leaks, reporting the findings) as this would involve a level of autonomy that is both overly complex and expensive. As a result, the uncertainty lies in how to create a process that is capable of delivering the same results that the manual process delivered while being faster, better and cheaper

No: If there is no change in process it is unlikely that a valid technological uncertainty exists as the core algorithm or task remains the same. 

Automation as a sub-activity of a larger project

So far, we have looked at two automation-based activities in isolation. However, in many cases the automation aspects of a project are simply a small cog in a large machine. This begs the question:

Are activities relating to the automation of a known process claimable under the R&D tax credit if the automation related activities are executed as a sub-activity of a larger project that would be qualifiable if viewed in a holistic sense?

The answer to this question can be found in Appendix 2 of the Revenue Tax and Duty Manual Part 29-02-05: Appointment of expert to assist in audits:

“…It will usually be appropriate to take a holistic view of a project rather than to focus on each component of the activities individually. Where it is felt that it is appropriate to break the activities down into smaller pieces, an explanation for the approach should be included”.

Let us take the above example of the Automation of the Functional Test plan; in isolation, the task of automating the Functional Test Plan does not qualify for the R&D tax credit due to the missing technological uncertainty and lack of technological advancement sought. Now, let’s consider if this task is a sub-activity of a project involving the design and development of a new big-data software analytics program (that from holistic standpoint satisfies all of the key criteria), the automation activity may be qualifying as it forms part of an overarching project.

As ever, it is important that a claimant company ensures they have considered the nature of all activities and how they may satisfy the key criteria for the R&D tax credit.

What can we learn from this?

When considering whether a particular automation-based project is likely to satisfy the requirements for the R&D tax credit, we must first ask The Process Question. From this we can draw the conclusion that if automation constitutes a sub-activity of a new or improved process, then the likelihood that it satisfies the criteria laid out by the legislation for qualifying activities increases. Conversely, if the activity involves solely the automation of a known process, then the likelihood of the criteria being satisfied decreases.

This article has been republished with kind permission from Irish Tech News.

Get in touch

The pace of change is challenging leaders like never before. To find out more about how KPMG perspectives and fresh thinking can help you focus on what’s next for your business or organisation, please get in touch with our team below. We’d be delighted to hear from you.