Artificial intelligence (AI) and robotic process automation (RPA) have the potential to significantly increase the efficiency of financial services organisations. Both use large data sets, which their algorithms and machine learning capabilities can process to aid decision-making.
Additionally, the compliance requirements of securities regulation require more and more data to be captured across multiple disparate communications channels.
What is the link between the two? How can new approaches to data capture, correlation, analysis and process automation turn tick-box compliance exercises into fuel for the future of finance?
In this paper, The Realization Group examines new models for generating smart business intelligence from captured communications data, using AI and RPA technology, with contributions from Will McDonald at Telstra; Bradley Hopkins of Company85; Steve Garrood of Insightful Technology; Ray Bricknell of Behind Every Cloud; Terry Walby at Thoughtonomy and Brian Wates of Pragmatic Process Automation.
Better than compliance
The story of compliance with the Markets in Financial Instruments Directive (MiFID II) should not end with a ticked box, but with a smarter, more efficient and self-aware business.
In its demand for managing communication, MiFID II has set a tremendous task for financial services firms, on both the buy- and sell-side. Their obligations under Article 76 of the Directive are to capture, store and monitor communications in such a way that they can be used to catch out bad actors.
The task is made tougher by the breadth of data, which encompasses everything from voice to electronic communication, and which must be assessed holistically in order to find patterns of behaviour that should be flagged up by compliance.
The implementation of a system that can pull all of this data together is one challenge. Managing and monitoring the flow and use of that data is another. Looming over both of those is the question, how can costs be managed effectively?
The answer, says Will McDonald, Business Development Director at Telstra, is in having the ability to repurpose compliance data for strategic business insight.
“A lot of firms will have some voice recording, certain surveillance tools and might look at other data that’s been captured,” he says. “The cool thing is when you’re able to gain strategic insight from integrating those functions across multiple systems and multiple forms of media.”
“The cool thing is when you’re able to gain strategic insight from integrating those functions across multiple systems and multiple forms of media.”
Will McDonald, Telstra
From a pure compliance perspective, the starting point for tackling Article 76 of MiFID II is the recording of telephone calls across landlines and mobile, text communications – whether via SMS, instant messaging or email – and other electronic communications. Storing that information is the first step, but more challenging is bringing it back together in such a way that regulators can understand any given trade within a reasonable timeframe.
McDonald gives an example: “One customer came to us because previously the regulator had given them a strict deadline, but physically they could not assemble all the data necessary within 72 hours. It just took too long to search, collate and store, because they had to provide it to the regulator in the format that they’d captured it in, with provenance of where it came from.”
Assuming data can be captured and stored, the real technical challenge is bringing that information together to form a single picture. This is made more difficult if the data is held across a patchwork of individual systems using multiple formats.
“Typically, the problem people have is that all this data is kept in different silos,” says McDonald. “We address that issue by putting all data in one discoverable place, converting unstructured data such as voice into searchable text, and then providing insight via modular AI-based tools and analytics packages, which opens up all sorts of opportunities.”
Radical change, without a revolution
To overcome the problem of silos, firms have two potential routes towards data aggregation. The first is to try to use separate point solutions for each function to capture information as and where trading takes place and then address the issue of data consolidation, normalisation and visualisation after the fact. The second is to take a more strategic view and deploy an integrated model that will capture, aggregate and deliver data as needed.
Although the second option creates many advantages, until recently investing in the installation of a comprehensive suite of software, communications tools and infrastructure to do this would typically require a considerable budget and a lengthy IT project. As a result, many firms have avoided such a strategic investment and let bottom-up IT investment define their firm’s architecture, for fear of over-committing both money and resources.
“The whole idea behind the cloud-based model is that it’s disruptive technology from a business value point of view, but it’s completely non-disruptive rom an implementation point of view.”
Terry Walby, Thoughtonomy
However, that fear can be allayed if cloud deployment of the tools and services is possible instead of an installed IT project.
Terry Walby, CEO of Thoughtonomy says, “The whole idea behind the cloud-based model is that it’s disruptive technology from a business value point of view, but it’s completely non-disruptive from an implementation point of view.”
Making better decisions
Onboarding a pre-integrated set of solutions-as-a-service not only overcomes the implementation challenges, it can turn data into a valuable asset, says Steve Garrood, Chief Commercial Officer at Insightful Technology.
“Forgetting the burden of regulation for a moment, what we are now seeing is a true digital transformation and use of real-time business intelligence within enterprises,” he says, “for example, where a bank can consolidate its entire communication base using a cloud-based environment, that ingests voice from IP telephony, turret or mobile, instant messaging and other e-comms at the back end to serve compliance, and then enables the use of that data at the front end to better serve the bank’s customers.”
Garrood explains that gaining a deeper and richer trend analysis of communications within the bank gives it a better viewpoint of its customer relationships, as the information can be scrutinised in order to find patterns that are supportive of other business objectives.
“This analysis can be circulated through their other associates or, indeed, other specialist areas within the bank,” he notes. “That’s taking the value of the data analysis from the compliance function and transferring it as actionable intelligence into the sales environment and customer experience.”
“All of these entities within the business are pushing for similar strategic goals, whether that’s to increase competitor advantage, get more insight from their customer, or reduce operating costs. ”
Bradley Hopkins, Company85
This model requires that data is not tied into siloes but made more easily accessible across functions, which means that a firm needs to take a strategic view of its objectives from front- to back-office and intelligently consider the ways in which the captured data can be applied to those objectives.
Bradley Hopkins, Head of Data and Analytics Consulting at Company85, says “All of these entities within the business are pushing for similar strategic goals, whether that’s to increase competitor advantage, get more insight from their customer, or reduce operating costs. However, certain functions will be able to drive more value from certain data sources or certain types of automation.”
Hopkins uses the term “data pragmatism” to describe the Company85 approach to such projects. “It’s a pragmatic approach to the world of data science, data engineering and data in general,” he says. “Business first, then approach, then technology.”
“It’s about process knowledge, not about technology knowledge. You’ve got the ability now, because of technology, to understand the process and understand the customer, and then decide what to do following that”
Brian Wates, Pragmatic Business Automation
This pragmatic approach also needs to be applied to process, argues Brian Wates, who recently left Fidelity Investments after 31 years to set up his own firm, Pragmatic Process Automation. “It’s about process knowledge, not about technology knowledge. You’ve got the ability now, because of technology, to understand the process and understand the customer, and then decide what to do following that. Then you’re able to get a new way of working,” he says.
Getting the right tools for the job
“We’re now in a world where it is realisable and achievable that real-time human interaction and communication of any form can be captured, converted to text, indexed, contextualised, correlated, presented in an accessible form either via a dashboard or via an API (application programming interface) and can be used to trigger activities and events,” says industry consultant Ray Bricknell.
“We’re now in a world where it is realisable and achievable that real-time human interaction and communication of any form can be… used to trigger activities and events.”
Ray Bricknell, Behind Every Cloud
But it doesn’t stop there. According to Walby, one of the advantages of having all this data available in a meaningful format is that robotic process automation (RPA) can be introduced to replace manual work and improve productivity. “RPA can be used to assist humans in doing some of the heavy-lifting work that’s part of day-to-day life, allowing them to focus on the things that they do better, or it can be used to execute work that might be unviable for a human to do, either because it is too labour-intensive or takes too long.”
“Why not make the technology required to meet compliance requirements the enabler for automating your processes and extracting true business value from your data?”
Steve Garrood, Insightful Technology
Ultimately, this approach is designed to provide a firm with not only comprehensive data compliance, but also a platform for gaining better insight across customer and supplier relations, while turning a cost into support for profit generation.
As Garrood says, “Why not make the technology required to meet compliance requirements the enabler for automating your processes and extracting true business value from your data?”