Case study
Eagle Eye: searching the web for early warning signals
AI case 5/16: early warnings for credit migrations
Imagine if you could search the entire internet to look for early warning signals for all kinds of events; to see whether a company is likely to experience financial distress in the near future, for example. This is exactly what Eagle Eye, Deloitte’s new and state-of-the-art AI tool, does.
Signals
Eagle Eye was developed by Deloitte Czech Republic. Analytics leader Jan Balatka and his team built a model to perform online semantic analytics in order to identify threats and opportunities. Balatka: “We initially made Eagle Eye for a financial company that wanted to know whether their creditors were likely to go into insolvency.” Traditional monitoring systems review creditors by checking their bank accounts, credit transfers or financial statements. But by the time you start to see warning signs there, it is too late, explains Balatka. “By then the company is already in financial distress.”
Before you can see it in the financial statements, early warning signals of potential decline can be observed online. This is precisely what Eagle Eye does. Balatka: “Eagle Eye uses open source intelligence to collect signals. It was built and tested for one particular problem, but the idea is to retrain the model for different markets and countries.”
How does Eagle Eye know what is relevant information – a signal – and what is not? Balatka: “It considers any and all information it finds about the company, client or market we assign it to as a signal.” With help from machine learning, Eagle Eye then starts to analyse signals, correlates them and recognises certain patterns. Balatka: “Only AI can handle the vast volumes of data on the internet and find correlations between parameters that humans would not even think of. Once we find certain patterns, Eagle Eye constantly monitors the internet to look out for them.”
Joint approach
The Czech team approached international colleagues to join their efforts and Roald Waaijer, Director Risk Advisory for Deloitte Netherlands, responded. Waaijer: “We were very interested in the technology. And when we tested the concept, our client feedback was really positive too.” It led to a collaboration with Balatka’s team and a joint approach to market.
Eagle Eye is tailor-made for each client, explains Waaijer. “Depending on their needs and preferences, clients can work with an application that gives them access to the monitoring system. Or we can simply supply alerts whenever a signal is spotted.” Balatka adds: “Eagle Eye serves as a starting point. Even though the system is right in the majority of the cases it flags, it still only provides an early warning. The client can follow up with its usual reviewing process.”
Waaijer sees a wide range of possible implementations of Eagle Eye: “Think of monitoring compliance, detecting fraud or being able to identify potential takeovers in a very early phase.” As of 2018, the technology is live and Eagle Eye is being used in different prototypes for clients across Europe.
*) 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
- 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
- PostNL: optimising delivery times
- 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.
Aanbevolen
SONAR: find labelling errors in databases
Case 2 out of 16 projects of applied AI