Automating the finance function

Automating the finance function

Mon 28 Sep 2020

Companies are introducing new technologies that are creating opportunities for finance departments. New self-service software has been in use since the summer of 2016 – but how can these companies implement robotic processes?

In less than five years, taxis have been ‘uberised’. Robot taxis — currently being tested in Japan — just like driverless Teslas, will in turn ‘uberise’ private chauffeur companies. New technologies are contributing to the emergence of new entrants to the market. This is true in our day-to-day life, and can also be seen in the world of business, as it undergoes a slow and steady, major (r)evolution.

New technologies have never been more accessible. The computing and storage capacities of IT tools are growing exponentially, particularly through processors that are becoming increasingly small.

The evolution of computer processing power in everyday life from 1956 to 2015:

In most companies, new business models continue to emerge, while organisations are shifting their strategies to refocus on their customers in new and different ways. Working methods are being reinvented – new generations are working differently and have new aspirations when it comes to the world of work, while jobs and careers continue to change.

Frequently, business applications are brought together in Financial Enterprise Resource Planning (ERP) systems which includes exchanges between stakeholders arranged in workflows. The complexity and cost of developments limit the improvement and adaptation of processes.

Robots designed for the finance function arrived in 2016. The name for the technology associated with these objects consists of three letters: RPA — Robotic Process Automation. These IT applications are implemented on companies’ IT servers or on the employee’s computers, mainly for the purposes of conducting repetitive or low-value tasks. These robots are automatic programs that are activated by operative events which are captured in the information system. These operative events may be recording a transaction in the ERP system, a threshold being exceeded, or a closing milestone.

The implementation of these robots is a response to the desire for operational efficiency.

The automatic programmes are generally configured by mimicking the actions of accountants, financial controllers and any other HR or IT system administrators. The specifications are created using a decision tree or a video recording of the movements of the mouse’s cursor onscreen. In a production environment, the IT application will carry out business transactions as often as is necessary, 24 hours per day, 7 days per week. By way of analogy, the robot transforms into a ‘virtual’ worker that can be quickly and easily trained to do the job required.
To qualify for automation, processes must fulfil two criteria: 

  1. The source data must be digital.
  2. The process structure must be stable and based on clearly established and formalised management rules.

Meeting these prerequisites does not appear to be too difficult for any mature organisation and the use cases are unlimited and can relate to all areas of finance. In practice, automated work is of transactional or clear decision-making nature. For example:  

  • Loading or aligning repositories. This is an example of a transactional case.After configuration, the robot is able to retrieve source data from a website (ie. a bank’s secured website) or the company’s information system (upstream in the process), collate this data into a spreadsheet and then integrate the data into a downstream application. The robot can finalise its work by sending a summary notification email to the accountant in charge of the process.
  • Managing discrepancies following the reconciliation of ‘intra-group’ or ‘intercompany’ transactions. This is an example of decision making using an algorithm. Usually, this type of reconciliation uses two data extractions from information systems that are specific to each legal structure (ie. the purchasing company and the selling company). In the event of a discrepancy, accountants from both companies coordinate with each other, adjust flows and provide rationale for their accounts.With RPA, extractions are edited and stored on the network. Discrepancies are calculated as defined in the specifications (ie. the calculation of discrepancies on the granular level of invoices). Entries are then directly recorded in the accounting system by the robot, on the basis of rules stated in the group’s closing instructions (ie, adjusting the intercompany flows for the seller’s position for all the flows that are of an amount above an indicated threshold). All the non-reconciled flows are treated as exceptions and may be listed in a spreadsheet and forwarded for manual processing.

Automation using RPA has three main benefits:

  1. Harmonising and securing processes.
  2. Reducing costs.
  3. Saving time for professionals who can focus on higher added-value work.

While these issues are easy to grasp, being organised is crucial when integrating the new technology into finance departments. Companies have generally organised themselves in four steps: training their teams, stating their requirements, experimenting and, finally, implementing within their finance departments.

The two first steps are standard but essential, especially in the case where a company is using new tools. Training increases employee awareness of technological developments and helps avoid potential obstacles involved in the transition. The statement of requirements is also an opportunity to ask questions, conduct initial calculations on the return on investment and measure the impacts for the organisation and information system.

Experimentation consists of testing the automation based on a simple use case (known as a Proof Of Concept (POC)), and then gradually expanding the scope of processes that will be automated in the test IT environment (known as the pilot), before migrating the pilot to the production environment in order to validate the estimated gains (project phase). Beyond the technical aspect of the organisation in project mode, the finance department will have to decide if it is a ‘go’ or ‘no go’ for each phase.

If a ‘go’ is given, the finance department will define the terms of use of this new RPA tool. Should it be applied across the board on all processes or is there a specific tool that can be used when needed by accountants and financial controllers?

If widespread deployment is chosen, the implementation of an RPA solution is gradually carried out process by process in order to keep producing figures. During this stage, the integration of robots is monitored in project mode to avoid a budgetary slide and in order to maintain the pace of deployment.

Many companies that have decided to integrate robots into their organisations have centralised their specification, development and maintenance skills within a dedicated team. In this case, the RPA platform must be managed like a centre of excellence in order to avoid being seen as a ‘black box’ — a processing apparatus where methods are opaque.

From the black box to the centre of excellence/centre of competence

RPA is a synonym for innovation in the finance function. It is highly important to understand how to use these new tools. The automation of processes has a ripple effect since it completely overhauls, standardises and improves manual processes. However, exceptions and deviations remain widespread despite a strong capacity for anticipating and controlling uncertainty. The added value of accountants and financial controllers will continue to shift towards being able to find solutions to difficult situations which cannot be solved by robots alone.