What will artificial intelligence mean for payroll?

What will artificial intelligence mean for payroll?
21 Sep 2017

Few words in the English language generate quite as much anxiety these days as automation, robotics and artificial intelligence - and concern appears particularly high among members of the payroll function. But are such worries justified?

 Most professionals’ uneasiness about automation stems at least partially from the fact that they fail to understand its benefits. These benefits include reducing or even eliminating time-consuming manual tasks so they can focus on higher-value activities, which ultimately make their jobs more interesting and their expertise more valuable to the organisation. Put another way, as payroll becomes increasingly automated, the profession has an opportunity to position itself as being of more strategic value. This shift has already taken place in many other functional areas such as sales, marketing and HR, making them more focused, productive and efficient and enabling them to deliver higher levels of performance.

But the ability of payroll to deliver strategic value to the business remains largely untapped.

As more and more global payroll organisations adopt cloud-based software and the function becomes increasingly data-driven, however, payroll information will become progressively important as a source of business intelligence.

Automating manual tasks

Broadly-speaking, automation, robotics, and artificial intelligence (AI) are on a spectrum ranging from simple to complex:

 

  1. Automation: This term refers to mechanising tasks and can be as simple as using mail merge to streamline document creation and mailings.
  2. Robotics: This concept is about using ‘bots’ to process, analyse and manage information. Its cousin, robotic process automation (RPA), is focused on controlling and automating rules-based processes so that they can be executed without human supervision.
  3. Artificial intelligence: These tools are highly sophisticated and are used to power systems such as IBM’s Watson, which employs humanlike thought processes to analyse and interpret wide-ranging data sets.

 

“As the adoption of RPA and AI grows, payroll professionals will see their roles evolve significantly, although the shift will not happen overnight.”

 When going down this route, the starting point, especially in payroll, is generally automation. As it rarely takes place across the board in any given function, process, or role, it is unlikely that the payroll department will ever be entirely automated. Instead, it is usually rote tasks or processes that are affected – those activities that are manual, repeatable, and adhere to a defined structure or set of guidelines.

 The sheer volume of these types of processes in payroll – combined with the massive amount of time and effort they absorb – mean it is surprising that so many of them are still manual. But following precedents in other functions, it would appear there are three particular payroll areas that are ripe for change:

 1. Data aggregation and validation:

Collecting, sorting, and sanitising the vast amount of information involved in payroll processing has traditionally required many hours of manual effort, much of which is spent passing data between human capital management systems, enterprise resource planning systems and payroll applications, usually via static Excel spreadsheets.

 This approach means that it becomes all too easy to load outdated, inaccurate, or incomplete data into systems for processing - unless payroll teams are prepared to spend hours combing and correcting it. Errors may be caught afterwards, but by that stage, the inaccuracies have already led to delays that take up, even more, staff time and effort to resolve.

 Automation can streamline data management in two ways though. Firstly, by using application programming interfaces (APIs), organisations can schedule automatic data transfers between their key administrative systems and payroll software in order to eliminate time spent on manual uploads and downloads.

 Secondly, pairing automated transfers with rules-based data validation systems means that payroll data can be cleansed automatically, both before and after processing. Doing so reduces the need for costly re-runs and results in more accurate data for analysis purposes and higher-quality reporting.

 2. Support calls and status updates:

Few enterprises have staff members dedicated exclusively to undertaking payroll support. The responsibility for answering queries and managing incident responses typically falls to payroll professionals, who already spend a lot of time each day verifying that everything is on track and prioritising tasks based on forthcoming deadlines.

 As a result, they often field dozens of calls a day from colleagues inside the organisation or from third-party vendors executing global payroll on their behalf. This situation can end up slowing pay cycles and increase the likelihood of human error. It is also where robotics and AI come in.

 For example, “chat-bots” today can be programmed to answer common payroll queries and escalate issues as required. Virtual assistants – similar in nature to Apple’s Siri or Amazon’s Alexa – will soon be available to streamline status monitoring and task management too. As a result, in future, you should be able to ask the system questions such as “What are my next five deadlines for payroll in the Europe, Middle East and Africa region?” and receive an instant response.

3. Process improvement and forecasting:

All too often, people believe that there is only one ‘right’ way to do things – that is, the way things have always been done. Even when looking for new systems, enterprises frequently focus on applications that can be configured to reflect their existing, and often rather convoluted, processes instead of taking the opportunity to transform their business for the better.

Another issue is that manual processes are impossible to capture, audit or measure in any systematic way. It is also very difficult to compare how they perform in comparison with others as part of a process improvement initiative. This means that many organisations have a blind spot around payroll performance and are unable to spot patterns or identify troubling trends.

But once payroll tasks are automated so too is the collection of related data for benchmarking and trend analysis purposes. By regularly analysing this information, payroll professionals can pinpoint possible areas where AI may be able to automatically “heal” processes by instantly identifying any data irregularities, recognising questionable patterns and, over time, automatically implementing new, corrective rules.

Preparing for the future of payroll

As the adoption of RPA and AI grows, payroll professionals will see their roles evolve significantly, although the shift will not happen overnight. As part of this process, the payroll team’s focus will increasingly shift to topics such as data integrity, workforce analysis, process improvement and performance management.

“In future, you should be able to ask the system questions such as “What are my next five deadlines for payroll in the Europe, Middle East and Africa region?” and receive an instant response.”

This means that now is the time for managers to prepare for change. Rather than fear automation, it should be seen as an opportunity to further develop your skills and understand where else you can apply your expertise in order to create value for the organisation. So to start the process, here are a few tips:

  1. Get more comfortable with data (and not just that held in Excel): One of the key components of any payroll automation project is a ‘unified data warehouse’ – that is, a single, consolidated source of accurate enterprise data. But before this is implemented, payroll professionals need to explore how payroll data currently moves across the organisation and at what points it is consolidated.

 

  1. Gain cross-functional knowledge and establish relationships: Many departments interact with payroll, including finance, HR, benefits administration, and compliance/ legal. Yet individual payroll professionals may not fully understand the resources that these functions require from them or how ultimately they are used. By building stronger relationships with colleagues in other departments, payroll practitioners can develop more cross-functional awareness and provide better services and support to departments relying on payroll information.

 

  1. Look beyond payroll to see the bigger enterprise picture: Focus on delivering as much value to the enterprise as possible by looking for new ways to adapt payroll rules and processes in a bid to achieve greater productivity, efficiency, and compliance. Also explore how the current wealth of payroll information can be turned into actionable information and used strategically within the wider business. For instance, by combining payroll data with information from finance, benefits and recruitment, payroll professionals can help the business to analyse pay trends more accurately in order to inform decisions around recruitment and organisational expansion.

 

What this all means is that, although automation in its various forms may appear on the surface to be anxiety-provoking, it could, strangely, do payroll a favour. By moving data to the heart of the equation and elevating the function’s significance beyond the tactical, it could just give payroll the leg up it needs to become an increasingly strategic partner for the business.

 

Brian Radin is president of CloudPay, a managed payroll system aimed at multinational businesses with complex global payroll requirements. He is responsible for all revenue-generating activities, strategic partnerships and business strategy at the company. Brian has more than two decades of leadership experience and a track record of boosting financial results for a number of leading human capital management companies. He has worked for Comcast, Booz & Co, and Aon Hewitt and also founded four start-ups. These include a global payroll software company, a professional employer organisation (PEO) and an employment and income technology platform supplier.

Few words in the English language generate quite as much anxiety these days as automation, robotics and artificial intelligence - and concern appears particularly high among members of the payroll function. But are such worries justified?

 Most professionals’ uneasiness about automation stems at least partially from the fact that they fail to understand its benefits. These benefits include reducing or even eliminating time-consuming manual tasks so they can focus on higher-value activities, which ultimately make their jobs more interesting and their expertise more valuable to the organisation. Put another way, as payroll becomes increasingly automated, the profession has an opportunity to position itself as being of more strategic value. This shift has already taken place in many other functional areas such as sales, marketing and HR, making them more focused, productive and efficient and enabling them to deliver higher levels of performance.

But the ability of payroll to deliver strategic value to the business remains largely untapped.

As more and more global payroll organisations adopt cloud-based software and the function becomes increasingly data-driven, however, payroll information will become progressively important as a source of business intelligence.

Automating manual tasks

Broadly-speaking, automation, robotics, and artificial intelligence (AI) are on a spectrum ranging from simple to complex:

 

  1. Automation: This term refers to mechanising tasks and can be as simple as using mail merge to streamline document creation and mailings.
  2. Robotics: This concept is about using ‘bots’ to process, analyse and manage information. Its cousin, robotic process automation (RPA), is focused on controlling and automating rules-based processes so that they can be executed without human supervision.
  3. Artificial intelligence: These tools are highly sophisticated and are used to power systems such as IBM’s Watson, which employs humanlike thought processes to analyse and interpret wide-ranging data sets.

 

“As the adoption of RPA and AI grows, payroll professionals will see their roles evolve significantly, although the shift will not happen overnight.”

 When going down this route, the starting point, especially in payroll, is generally automation. As it rarely takes place across the board in any given function, process, or role, it is unlikely that the payroll department will ever be entirely automated. Instead, it is usually rote tasks or processes that are affected – those activities that are manual, repeatable, and adhere to a defined structure or set of guidelines.

 The sheer volume of these types of processes in payroll – combined with the massive amount of time and effort they absorb – mean it is surprising that so many of them are still manual. But following precedents in other functions, it would appear there are three particular payroll areas that are ripe for change:

 1. Data aggregation and validation:

Collecting, sorting, and sanitising the vast amount of information involved in payroll processing has traditionally required many hours of manual effort, much of which is spent passing data between human capital management systems, enterprise resource planning systems and payroll applications, usually via static Excel spreadsheets.

 This approach means that it becomes all too easy to load outdated, inaccurate, or incomplete data into systems for processing - unless payroll teams are prepared to spend hours combing and correcting it. Errors may be caught afterwards, but by that stage, the inaccuracies have already led to delays that take up, even more, staff time and effort to resolve.

 Automation can streamline data management in two ways though. Firstly, by using application programming interfaces (APIs), organisations can schedule automatic data transfers between their key administrative systems and payroll software in order to eliminate time spent on manual uploads and downloads.

 Secondly, pairing automated transfers with rules-based data validation systems means that payroll data can be cleansed automatically, both before and after processing. Doing so reduces the need for costly re-runs and results in more accurate data for analysis purposes and higher-quality reporting.

 2. Support calls and status updates:

Few enterprises have staff members dedicated exclusively to undertaking payroll support. The responsibility for answering queries and managing incident responses typically falls to payroll professionals, who already spend a lot of time each day verifying that everything is on track and prioritising tasks based on forthcoming deadlines.

 As a result, they often field dozens of calls a day from colleagues inside the organisation or from third-party vendors executing global payroll on their behalf. This situation can end up slowing pay cycles and increase the likelihood of human error. It is also where robotics and AI come in.

 For example, “chat-bots” today can be programmed to answer common payroll queries and escalate issues as required. Virtual assistants – similar in nature to Apple’s Siri or Amazon’s Alexa – will soon be available to streamline status monitoring and task management too. As a result, in future, you should be able to ask the system questions such as “What are my next five deadlines for payroll in the Europe, Middle East and Africa region?” and receive an instant response.

3. Process improvement and forecasting:

All too often, people believe that there is only one ‘right’ way to do things – that is, the way things have always been done. Even when looking for new systems, enterprises frequently focus on applications that can be configured to reflect their existing, and often rather convoluted, processes instead of taking the opportunity to transform their business for the better.

Another issue is that manual processes are impossible to capture, audit or measure in any systematic way. It is also very difficult to compare how they perform in comparison with others as part of a process improvement initiative. This means that many organisations have a blind spot around payroll performance and are unable to spot patterns or identify troubling trends.

But once payroll tasks are automated so too is the collection of related data for benchmarking and trend analysis purposes. By regularly analysing this information, payroll professionals can pinpoint possible areas where AI may be able to automatically “heal” processes by instantly identifying any data irregularities, recognising questionable patterns and, over time, automatically implementing new, corrective rules.

Preparing for the future of payroll

As the adoption of RPA and AI grows, payroll professionals will see their roles evolve significantly, although the shift will not happen overnight. As part of this process, the payroll team’s focus will increasingly shift to topics such as data integrity, workforce analysis, process improvement and performance management.

“In future, you should be able to ask the system questions such as “What are my next five deadlines for payroll in the Europe, Middle East and Africa region?” and receive an instant response.”

This means that now is the time for managers to prepare for change. Rather than fear automation, it should be seen as an opportunity to further develop your skills and understand where else you can apply your expertise in order to create value for the organisation. So to start the process, here are a few tips:

  1. Get more comfortable with data (and not just that held in Excel): One of the key components of any payroll automation project is a ‘unified data warehouse’ – that is, a single, consolidated source of accurate enterprise data. But before this is implemented, payroll professionals need to explore how payroll data currently moves across the organisation and at what points it is consolidated.

 

  1. Gain cross-functional knowledge and establish relationships: Many departments interact with payroll, including finance, HR, benefits administration, and compliance/ legal. Yet individual payroll professionals may not fully understand the resources that these functions require from them or how ultimately they are used. By building stronger relationships with colleagues in other departments, payroll practitioners can develop more cross-functional awareness and provide better services and support to departments relying on payroll information.

 

  1. Look beyond payroll to see the bigger enterprise picture: Focus on delivering as much value to the enterprise as possible by looking for new ways to adapt payroll rules and processes in a bid to achieve greater productivity, efficiency, and compliance. Also explore how the current wealth of payroll information can be turned into actionable information and used strategically within the wider business. For instance, by combining payroll data with information from finance, benefits and recruitment, payroll professionals can help the business to analyse pay trends more accurately in order to inform decisions around recruitment and organisational expansion.

 

What this all means is that, although automation in its various forms may appear on the surface to be anxiety-provoking, it could, strangely, do payroll a favour. By moving data to the heart of the equation and elevating the function’s significance beyond the tactical, it could just give payroll the leg up it needs to become an increasingly strategic partner for the business.

 

Brian Radin is president of CloudPay, a managed payroll system aimed at multinational businesses with complex global payroll requirements. He is responsible for all revenue-generating activities, strategic partnerships and business strategy at the company. Brian has more than two decades of leadership experience and a track record of boosting financial results for a number of leading human capital management companies. He has worked for Comcast, Booz & Co, and Aon Hewitt and also founded four start-ups. These include a global payroll software company, a professional employer organisation (PEO) and an employment and income technology platform supplier.

Leave a Reply

All blog comments are checked prior to publishing