Finance
How BFSI Use Natural Language AI and RPA for Efficiency and Accuracy in Financial Reporting and Fund Commentary
By Mark Goodey, Managing Director & Innovation Strategist, Arria NLG
It has been nearly a century since John Maynard Keynes predicted that by 2030, the average workweek would be compressed into a mere 15 hours, turning the traditional nine-to-five into a breezy nine-to-12. Keynes prediction has obviously not materialised, and the current average isn’t trending in the right direction. In finance, it has reached a point that some have begun politicking for an 80-hour cap to account for sleep and other human needs. However, the accelerating adoption of new technologies – from robotic process automation (RPA) to machine learning, natural language technologies (NLT) and other AI-fueled innovations – means these advances are finally yielding efficiencies that don’t simply minimise costs, but also free up staff through “end-to-end” automation requiring no human intervention.
The benefits of these new capabilities are not always obvious. For example, investors may not realise the shareholder communications they study every quarter did not require a human analyst to put pen to paper; instead, it is possible that NLG automated the investment commentary. Furthermore, technology’s reliance on acronyms and jargon doesn’t help operators understand the value.
According to Forrester, NLG can reduce the manual labour of report generation by 80 percent, saving up to 60 percent in time spent by Business intelligence (BI) analysts and their supported decision-makers.
Further, intelligent automation removes any risk of human error in reporting and written analyses, increasing annual revenue up to 2 percent. NLG produces data-driven narratives that ensure reporting accuracy and improve operational efficiencies.
For fund managers and financial advisors, Forrester estimates that natural language AI reduces the manual resources required for written commentary by as much as 85 percent.
RPA, for the uninitiated, is software that builds, deploys, and manages artificial intelligence (AI) that interacts with digital systems and software. The technology can understand what’s on a screen, complete keystrokes, navigate systems, identify and extract data, and perform a wide range of defined actions. Natural language generation (NLG), meanwhile, is exactly what the name implies. It’s a subset of AI that transforms data into natural language, replicating the human process of analysing and communicating data-driven insights.
Nearly any decision in finance can be traced back to three primary objectives: save money, make money, or manage risk. The combination of RPA and NLG supports these goals. Both RPA and natural language AI deliver material efficiencies. While finance may slow to adopt new technologies – at least compared to other industries – the sector is also uniquely positioned to realise the greatest benefits.
When RPA and natural language technologies are joined in a single platform, it is a game-changing innovation that impacts the entire organisation.
Contextualising the Pairing
How do the two technologies work together?
RPA software extracts and assembles both structured and unstructured data from multiple sources. Then the NLG platform creates written reports instantaneously based on the data. RPA can format the reports and distribute PDFs to appropriate constituencies.
In financial settings, RPA technology eliminates manual tasks around performance measurement and attribution to automatically organise and analyse return data. NLG handles middle-office roles, such as drafting, iterating, and reviewing copy and commentary. Beyond eliminating human errors, “Intelligent Automation” also improves compliance, accelerates performance, facilitates better access to insights, and improves data literacy in the process.Savings and efficiencies represent the most obvious payoff. For instance, an overlooked challenge only back- and middle-office professionals will appreciate is that there is no “load-balancing” for reporting on monthly, quarterly or annual schedules. Historically, adding new funds and products requires significant investments in the back- and middle-office to keep up with demands, which have only become more onerous in today’s evolving regulatory landscape. When RPA and NLG technologies come together, investment managers can suddenly scale far more efficiently, adding new assets without investing in support staff.
Another benefit is that the intelligent automation doesn’t become any more expensive as assets under management (AUM) grows.
NLG-authored commentaries and fund reports are indistinguishable from traditional copy produced the old-fashioned way because the written analyses can be customized to reflect context and the voice of the firm. For asset managers and corporate finance teams, utilising narrative models based on asset class can customise the vernacular – but never alter the facts. Randomisation functions make it impossible for even the most vigilant readers to discern patterns from one commentary to the next. The ability to stay in front of key audiences – staying top of mind – supports marketing and client retention.
Perhaps the biggest advantage from the perspective of compliance professionals is the extent to which RPA and NLG technologies reduce risk. Errors are common across any industry that deals with data, but they are can be exceedingly costly when they occur in finance and investment management. For example, the 2010 “flash crash” was initially attributed to human error. By automating data extraction, organisation, and analysis, asset managers eliminate such risks.
Another factor is that these technologies eliminate human biases. Regulators will cite inconsistent disclosures, misleading statements, and omissions of material information as red flags deserving scrutiny. The “indifference” of technology imparts a sense of trust.
One of the challenges limiting adoption of RPA and NLG technology is that these two capabilities are also two of the best kept secrets: RPA is already ubiquitous across back- and middle-office workflows, but casual observers don’t have a window into what’s being automated; NLG, alternatively, is designed to be indistinguishable from commentary written by humans and it is prevalent across many industries, including news and media. In finance, it is more common one may think, deployed across everything from P&L reporting to ESG analysis and attribution, credit ratings, and other applications.
McKinsey identified RPA and NLG as two of the five core technologies that will power the next generation operating model. Meanwhile, Forrester quantified that the ROI from these two technologies will pay for themselves in less than a year.
Although NLG may not translate into Keynes promised 15-hour work week, it does alleviate humans from some of the drudgery involved in manual reporting.
Most importantly, because it translates commentary directly from data, it is 100% accurate and unbiased.
About Author:
As Managing Director and Innovation Strategist for Arria, Mark Goodey leads the company’s Investment Analyst practice in Banking, Financial Services, and Insurance (BFSI). Prior to joining Arria, Goodey served as lead for BNY Mellon’s Data & Analytics Solutions where he worked to establish Eagle Performance as a leader in investment performance analytics. He was also a prior member of several organizations and governing bodies that include UK Investment Performance Committee (UKIPC); Performance Measurement and Client Reporting Review; The Journal of Performance Measurement; Performance Measurement Networking Group; and Barclays Capital Index Advisory Council.
This is a Sponsored Feature.
-
Top Stories4 days ago
BoE’s Bailey says getting inflation to 2% will be ‘hard work’
-
Top Stories4 days ago
Marketmind: Markets turn risk-averse after bumper month
-
Top Stories4 days ago
Baer says exposure to single group tops 600 million francs, as Signa crisis deepens
-
Top Stories4 days ago
UK’s FTSE 100 dips on miners, energy drag