Argus: AI for document interrogation and analysis Bookmark has been added
Argus: AI for document interrogation and analysis
Case 15/16 of applied Artificial Intelligence
The Engagement team at Deloitte doesn’t need to be told how massively labour-intensive it is to scan through lengthy contracts in order to identify minor differences. Fortunately, they now have a helping hand in the form of Argus, an Artificial Intelligence solution that compares documents quickly and picks out the tiniest details and discrepancies.
“Our department assesses an awful lot of documents”, says Twan van Gool, Director Innovation & Analytics within the Audit department at Deloitte. “Take referral instructions, for example: internal agreements requiring a number of confirmations from Deloitte auditors in other countries.” It is often the case that one of these instructions is returned with a minor amendment. “The original states that the audit is to be carried out in English, but in the returned version, it has suddenly changed to Russian or Portuguese,” Van Gool continues. “By doing this, a team is trying to fulfil specific wishes while keeping an eye on their own budget.”
It is a burdensome and time-consuming task to spot these kinds of minor amendments. “I have seen people placing one page on top of the other and holding it against the light in an attempt to quickly identify whether anything has changed. That’s obviously not the most effective way of going about it,” remarks Van Gool. He is therefore very happy with the introduction of Argus, a smart tool that scans and analyses documents using cognitive technologies.
“First of all, you teach Argus the structure of a contract, and what the components are that are permitted to vary from contract to contract, such as the name, place, and date. Argus is then able to make comparisons and search for differences,” he explains. Argus analyses the differences, assigns the amendments to categories according to how significant they are, and delivers a risk report.
AI email alert
Get 16 cases of applied AI in your mailboxSign up here
Argus was being used at first within the Audit departments of Deloitte US, Canada and Australia. In September 2018 Argus was introduced in The Netherlands as well. The results of using Argus are plain to see. Departments that are already using it have noticed that reviews of contracts are being carried out far more efficiently and with a higher level of quality.
At first, the tool is expected to be used mainly for analysing differences between contracts. “But there are unprecedented opportunities for providing support in the analysis, appraisal and comparison of documents,” notes Van Gool. “And we are rapidly developing algorithms for analysing other types of documents, such as invoices.” For example, he believes that Argus will bring about an end to manually searching for details in labour-intensive, data-orientated audit tasks, “such as manually re-entering the total amount from a scanned invoice. Very soon, Argus will do that for you.”
The Argus licences are valid specifically for Audit, but Van Gool also recognises opportunities to use it in other Deloitte departments. “Argus could make a major difference in Consulting or Tax. In fact, anywhere where more than a handful of contracts need to be compared, or information needs to be copied across from documents.”
*) This case is part of the series of 16 Artificial Intelligence projects from Deloitte. Other cases in the series are in random order:
- TAX-I: A virtual legal research assistant
- AI Benchmark for transfer pricing
- SONAR: Find labelling errors in databases
- Transaction detector with regard to the Dutch work cost regulations
- GRAPA: assistance with risk strategies
- Chatbot as a handy search tool for the online technical library
- Argus: an eye for detail
- Virtual assistants: beyond the hype
- HR agent Edgy: the future of Human Resources
- Using machine learning to assess risks for insurance policies
- Predicting payment behaviour
- DocQMiner: contract analysis performed in no time at all
- Combating welfare fraud with machine learning
- Using machine learning and network analytics to search for a needle in a haystack
- Clustering unstructured information in BrainSpace
Sign up for the email alert to get all cases through email.