GRAPA: assistance with risk strategies

Case studies

GRAPA: assistance with risk strategies

AI case 10/16 of applied artificial intelligence

When auditors determine a risk strategy, they partly base it on knowledge that they gained in previous audits. Deloitte is now developing a smart personal assistant that supports auditors using the pooled expertise of all their fellow professionals.

Knowledge gained

Determining the risk strategy forms an important part of an audit. Twan van Gool, Director Innovation & Analytics within the Audit department at Deloitte, explains: “The auditor determines which sections of the annual accounts are high-risk, based on market developments, new legislation and regulations or incidents within the company, for example. The chosen risk strategy determines the subsequent audit method.”

When formulating the risk strategy, an auditor’s knowledge gained from previous cases will be very valuable. Van Gool: “Every business is different, of course, but an audit is an audit,” notes van Gool. “While lingerie and cycling might have nothing in common, the retail and stock process of Hunkemöller is largely the same as that of Halfords. So the associated risks are also similar to one another.”

AI email alert

Receive the latest AI cases


A second reader

Imagine how useful it would be if auditors could use not only their own knowledge and experience, but also those of fellow auditors, when determining the risk strategy. With that idea in mind, van Gool’s team developed the AI tool Guided Risk Assessment Personal Assistant, or GRAPA for short.

“GRAPA assists an auditor in marking out their chosen strategy against all other risk strategies that have been used before,” explains van Gool. It uses a Deloitte database of 10,000 cases, and each case contains an average of fifty risks.

GRAPA is not a stand-alone application; rather, it is added to the software that auditors use when determining the risk strategy. Van Gool: “It’s as if you can ask a second person to read alongside you,” van Gool continues. “But the advantage is that this second reader has the pooled expertise of Deloitte.” He emphasises that the auditor remains responsible for the chosen risk strategy and audit method. “GRAPA indicates what has happened in similar cases. But if there is anything about the client’s situation that is special or unusual, it is obviously down to the auditor to tailor the approach accordingly.”

Benchmarking and planning

The tool is therefore not intended to replace auditors. “When it comes to critical consideration of processes, developments and risks, you still need creativity and human intelligence,” maintains van Gool. However, he does recognise there are opportunities to automate the completion of an audit once the risk strategy and audit method have been determined: “Those activities are so standardised that a robot could do them.”

Deloitte intends to use GRAPA in 2019 for the audit of the 2018 tax year. “We are currently working on a proof of concept,” says van Gool. The next step is to use GRAPA for more than just examining data from the past. “We want to feed the tool with market development data, such as those relating to Brexit, or high-risk derivatives in the housing market. That knowledge will enable the tool to also present potential future risks to the accountant.”

*) 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.

Did you find this useful?