Transaction detector with regard to the Dutch work cost regulations

Case study

Transaction detector with regard to the Dutch work cost regulations

AI case 7/16: WKR analytics

The work-related expenses scheme in the Netherlands, known as WKR, causes a world of problems for many companies. Deloitte has developed a clever solution to assist its clients in implementing this tax scheme correctly.

Unfamiliar scheme

The WKR permits companies to spend 1.2 per cent of taxable wages on tax-free allowances and benefits in kind for their employees. “When the scheme became compulsory for all employers back in 2015, it soon revealed a number of problems,” explains Guy Thien, tax consultant at Deloitte. “In practice, for example, there is a major lack of clarity regarding who is responsible for implementing the scheme. It requires data from HR, salary and financial records, to name a few, but no one takes the role of managing it.”

What is more, it is still a fairly unfamiliar scheme, and consequently many expenses are incorrectly excluded from it. Thien: “If a director gives ten employees an iPad each, he is often unaware that this has implications under the WKR,” continues Thien. “He fails to inform the appropriate parties, and the expense falls through the cracks.”

AI email alert

Receive the latest AI cases

Sign-up

Categorising and learning

In order to overcome these problems, Deloitte’s Financial HR Analytics team developed an intelligent system that reads descriptions of expenses and is able to categorise them: WKR Analytics. “You enter all your expenses, and the system then picks out what is covered by the WKR. Without our tool, a client might achieve a quality level of 20 per cent in its implementation of WKR. We increase it to 95 per cent,” asserts Thien.

A tax expert is called in for the remaining 5 per cent. The expert wastes far less time on WKR than was previously the case, and is able to focus his attention on the doubtful cases. Thien: “If WKR Analytics categorises an expense incorrectly, you can correct it and the system will learn from it,” explains Thien. “Every correction makes the system smarter and more efficient. The longer you use it, the less time you waste on it.”

WKR Analytics comprises a combination of machine learning technologies and natural language processing – in other words, knowledge of language and use of words. Thien: “The WKR is relevant in specific situations,” continues Thien. “Take going for lunch. When I go and have lunch with a colleague, that may have different implications under the WKR to when I go for lunch with a client. WKR Analytics understands that distinction.”

The advantage

WKR Analytics is now being used by over forty clients. Thien: “I have noticed that the time is ripe for smart solutions,” says Thien. “Back in 2014, when I told employers about WKR Analytics, they scarcely believed that it would really work. Now, they are actually curious as to what we can offer.” Not only is the number of users growing, WKR Analytics is continuing to develop too. “We are continuing to improve the dashboard, as well as adding new information and constantly increasing the degree of integration with VAT analytics,” says Thien. “We are now also demonstrating what the level of data quality is. In just two hours, we can know whether a company’s data are good enough to enable WKR Analytics to be used.”

The success of WKR Analytics has not gone unnoticed. There are even a few companies that have attempted to copy it. Thien considers it a compliment: “That doesn’t scare me. “Once you have gained an advantage with AI, you don’t lose it very quickly.”

*) This case is part of the series of 16 Artificial Intelligence projects from Deloitte. Other cases in the series are in random order:

  1. TAX-I: A virtual legal research assistant
  2. AI Benchmark 
  3. SONAR: Find labelling errors in databases
  4. Transaction detector with regard to the Dutch work cost regulations
  5. GRAPA: assistance with risk strategies
  6. Chatbot as a handy search tool for the online technical library
  7. Argus: an eye for detail
  8. PostNL: optimising delivery times
  9. Virtual assistants: beyond the hype
  10. HR agent Edgy: the future of Human Resources
  11. Using machine learning to assess risks for insurance policies
  12. Predicting payment behaviour
  13. DocQMiner: contract analysis performed in no time at all
  14. Combating welfare fraud with machine learning
  15. Using machine learning and network analytics to search for a needle in a haystack
  16. Clustering unstructured information in BrainSpace

Sign up for the email alert to get all cases through email.

Vond u dit nuttig?