Smarter Humans. Smarter Machines.

Patrick Donaldson is Head of Market Development, Wealth Management for Asia Pacific & Japan at Refinitiv, the USD6 billion plus revenue global provider of financial markets data and infrastructure. He presented at the Hubbis Digital Wealth Management Forum to highlight to delegates how AI will be the single greatest enabler of competitive advantage in the financial services sector, but that data quality is the biggest barrier to the adoption and deployment of machine learning.

Refinitiv is a global provider of financial markets data and infrastructure. The company was founded in 2018. It is jointly owned by Blackstone Group LP which has a 55% stake and Thomson Reuters which owns 45%. The company has an annual turnover of USD6bn with more than 40,000 customers in 190 countries.

“There are a few inconvenient truths that we need to address within the wealth management industry around us,” he began. “The first is that for all these digital solutions to work, the data that goes into them has got to be world class. And the other truth is that if you in the industry do not recognise this and fix it, you are falling behind your competition.”

Play, or don’t compete

Donaldson referred to the recent Refinitiv 2019 Artificial Intelligence/Machine Learning (AI/ML) Global Study amongst the larger financial institutions. “We saw that some organisations are way ahead of where we had expected and are making machine learning a core component of their business strategy, with growing investment in this area. The leaders are in the US and they are the largest players. We in the wealth industry must follow their lead, or risk being left behind. A few years ago, if you were a compliance officer you could almost name your price to get a job in the wealth management industry. Nowadays the data scientists are naming their price.”

The survey of data scientists

The Refinitiv AI/ML survey was assembled from 447 interviews and discussions with data scientists and technology experts in financial institutions with more than USD1bn in revenue.

“The commitment to implementing AI/ML in our market is further advanced than we had expected,” Donaldson explained, “with over 90% of those surveyed having deployed ML to manage or analyse content to one or more departments in their organisation. And 78% state that ML is a core component of their business strategy, driven by better information, better insights and greater productivity, but not by motivations surrounding cost cutting.”

Seeking quality

Data discoverability and quality are the biggest barriers to the adoption of AI/ML. “Some 43% stated poor data quality impacts their ability to adopt ML, while data scientists spend a high proportion of their time ‘wrangling’ the data for use in ML. There is such a massive volume of financial data with diversity in the structure and the source, and accordingly managing this data is a significant challenge. One of the most complicated tasks is to get data which is relevant, reliable and from a secure source which has some statistical value.”

Connecting the dots

The key challenge, he explained, is capturing multiple different data sources, then linking all of it together, then standardising it all.

“If well organised and executed,” Donaldson observed, “what that means for wealth management is driving actionable relevant insight or – to put it another way – giving the right advice to the right client at the right time. After all, an individual relationship manager or an investment adviser could only offer so much advice to so many clients during their average day. We just have to automate the process and start pushing that information directly to the client, thereby massively increasing the productivity and performance, as well as ramping up the quality of the end user experience.”

Donaldson then pointed to what he called the biggest challenge that emerged from the research. “The problem is for the data scientists actually getting their hands on the data, yet alone scrubbing it, standardising it or even thinking about how they can analyse it. Getting hold of the data is the core issue.”

He noted also that the types of data that are being requested are also changing. “Within wealth management, you have access to a large amount of personal information about your clients, but not much of it is necessarily unique as your competitors probably have largely the same information. Furthermore, your clients may be sharing much of their personal information on social media sites much of the time.”

Donaldson highlighted how the vital mission is to extract data available within the wealth management organisations that the clients share and that is available through multiple other sources. “This can make the difference in your business between giving a good service and giving a great service,” he stated. “Personal, relevant and timely insights driven by AI greatly improve productivity, so address this issue urgently, or you will be left behind by your competitors.”

Your competitive edge

“We expect there to be an explosion in the use of AI/ML across financial services, enabled by new tools running in the cloud,” Donaldson summarised on closing his talk. “Such tools will reduce the need for all firms to build large dedicated data science teams. But customers need to connect their data and leverage available ontologies. Unstructured data is key, and mastery of Natural Language Processing (NLP) is key, as is overcoming the staffing and skills shortage amongst the data scientists.