Random checks are still perfectly common in the financial sector. Still, that feels like a weakness for Crowe Foederer’s data-driven accountants. They do not find taking samples and demanding valuable time from customers very service-oriented and the quality of service can be much better by taking care of the entire data. The goal is to let all the numbers pass through their hands. But automated.
Functionally, Crowe Foederer had already worked out her dream image: you pick all relevant data from the customer’s ERP system, you load it into your own system and dashboards with advanced insights are made available to the customer. Pipple was the right party to enrich Crowe Foederer’s self-developed data science platform with a modern data warehouse, judged Arjen van Zon, Information Officer at Crowe Foederer. “They know our cloud platform Microsoft Azure like the back of their hand, they are proficient in Python and they teach a team the principles of design thinking as they go.”
Safe and transparent
The auditors presented Pipple with roughly three challenges. The connection had to be ultra-secure. After all, it is quite a lot for a customer to make his data available. There also had to be insight into the quality of the data, so that Crowe Foederer could also advise its customers in this area. Finally, the automatic analysis had to be able to be rewound in slow motion to the original data, in order to be able to justify the audit results to the supervisor in a crystal clear way. In short, a complete quality solution that fits the data science platform of Crowe Foederer.
For this purpose, Pipple built a so-called ‘data lake’: a next generation data warehouse, but a lot more advanced. In addition to structured data, you can also store unstructured data in it. Images, documents, videos and audio files, for example. To fill the data lake, Pipple built a kind of car wash in which the data goes through three phases. In the first phase, the customer data is copied one-on-one from the customer’s system and stored in the cloud. There, all data is checked: isn’t it corrupt, is it complete? In phase two, the quality of the data is checked: for example, is the formatting correct, and are the values realistic? In the final stage, the entire file is converted to the format that Crowe Foederer itself uses. the Common Data Format, so that the dashboards can be generated and published for use.
Inventing the wheel together
Pipple and Crowe Foederer spent almost a year developing it together. That process went smoothly, arjen thinks. ‘We have really created something innovative together. Even large companies in Insurance and Banking do not do this automatically. So we had to invent the wheel together. For that you have to be critical and that was possible: Pipple never shot in the defensive.’. He likes the fact that Pipple is a relatively small agency. ‘At larger firms, the collaboration is immediately formal: they encapsulate everything legally and set up a heavy project organisation. That makes the process slower and more expensive. Pipple is accessible and hands-on with pleasant people to work with.’
Crowe Foederer takes care of the further development for herself, but knows how to find Pipple again in the future: ‘In fact, we have had Pipple create a basic structure on which our people can continue to develop. Pipple trained them in that. But I don’t rule out using them in the data intelligence field.’