In response to the many constraints faced by banks and financial institutions, Scaled Risk offers a range of solutions based on the rapid and flexible use of large volumes of data in real time. Scaled Risk's platform, based on the exploitation of Big Data, provides business lines with unprecedented speed and agility.
"Scaled Risk's objective is to bring data closer to the business user. Both to make their lives easier, but also to make them independent of IT", explains Hervé Bonazzi, CEO of Scaled Risk.
Bringing the business closer to the data, whatever the volume to be processed. Since its creation in 2012, Scaled Risk has understood the potential of Big Data technologies to support the digital transformation of banks and financial centres. To achieve this, the start-up is developing a data management platform that complements the Hadoop technology base to provide a solid, scalable and flexible transactional system, and above all provides data auditing and traceability capabilities. These are essential functions for meeting the regulatory constraints inherent in the financial sector.
A data lake for analysing and tracking data
Based on this platform, Scaled Risk makes it possible to build a lake of data from multi-channel data collection. Within this data lake, the data models are dynamic, flexible and, above all, 'versionable'.
Thanks to a simple modelling interface, this data can be exploited for a wide range of uses, as Hervé Bonazzi explains: "The strength of our platform is that it provides a logical layer that enables the various business units to model their data in order to build their calculations. We provide an on-demand analysis engine, a sort of 'analytical self-service'.
With this tool, business users have simple tools for manipulating their data, reconciling it and drawing up their own reports and analyses. Depending on their business, operators will be able to carry out risk analysis, draw up their P&L, respond rapidly to regulatory reporting requirements, and analyse customer behaviour, among other possible uses. Regulatory reporting is a good example for Hervé Bonazzi. These reports are increasingly frequent and exhaustive, and therefore require more data. By consolidating the data in a data lake and querying it with the same tool, we gain in speed compared with queries on numerous databases.
"With this system, the bank can have a very deep history on which we use classification and machine learning algorithms (see below), which enables us, for example, to detect patterns of fraud or customer behaviour (investor type) thanks to the detection of signals. It's this logic of storage and platform agility that gives the financial institution the flexibility it needs."
The cloud as a guarantee of agility
In fact, banks are looking for this flexibility and agility to face up to digital transformation and resist the more nimble fintechs with sharpened business models. On this last point, Scaled Risk's SaaS model (hosted securely and privately by Cloud Temple) was created from the outset to adapt to use cases and offer the elasticity needed for digital transformation.
For Hervé Bonazzi, "the banking industry needs to undergo a digital transformation in order to face up to the fintechs who are arriving with new models. At present, many financial players have very little in the way of a modern, digital relationship with their customers. The Mifid II regulations impose and provoke this digitisation, so everyone will have to carry out their digital transformation, and that's where we are positioned," concludes Hervé Bonazzi.
Artificial intelligence is that of the Data Scientist
Among the many uses of Scaled Risk, the fight against fraud is a key argument. "It's a major issue. Compliance is essential because banks are fined for dealing with prohibited or fraudulent counterparties. The regulator is increasingly coercive with regard to money laundering. The use of machine learning, with a real-time event engine, makes it possible to reduce false positives by raising an alert if there is any doubt about an interaction. At present, banks have too many false positives to deal with by human means. To reduce false positives, we need to teach the machine to refine its algorithms".
For Hervé Bonazzi, there are no miracles and no superior artificial intelligence. To be efficient, you have to constantly learn from the machine. The real intelligence lies with the data scientist and his or her ability to adapt the algorithm to the business. "The algorithms we use are old. But today we have the data depth to perform classification, detect learning with a feedback loop that refines the classifications according to the data fed back. The efficiency of artificial intelligence lies in its ability to adapt to your needs and uses as a function of your inputs. This adaptation is based on the system's dynamic ability to tend towards a point of convergence between the right compromise and the right limit. This enables it to meet the customer's needs".
Listed by Usine Nouvelle, the SCALED RISK solution enables :
- collect structured and unstructured data in real time from heterogeneous systems,
- organise this data in a flexible, scalable model,
- immediate access thanks to an intuitive search engine,
- build real-time analytics in a real-time, distributed, in-memory OLAP cube,
- publish via API or visualise data along an infinite number of axes,
- vary the time axes associated with the data,
- compare, republish and alert with sub-second response times,
- produce audit evidence for the regulator.
Scaled Risk can be deployed on the customer's infrastructure or accessed as a SaaS solution, and is fully open via APIs (REST, Java, Excel). Scaled Risk natively integrates the FpML format.