When you have the leading names in data management gathered in one New York City hall, it would be remiss not to ask them how they view some of the most pressing issues facing the industry today.
So, that’s exactly what A-Team Group did at our 14th annual Data Management Summit in the city. We seized the opportunity to take the temperature of the global data industry by posing a series of polling questions at the event that touched on the topics that matter most.
Their responses were often surprising, sometimes predictable but always fascinating. In the first of two reports, we look at the discussion points raised on the application of artificial intelligence to data management, data modernisation strategies and the use of data products. Next week, we’ll examine attitudes to regulatory reporting data, self-service products and new tooling challenges.What do you consider to be the biggest challenges in realizing ROI from adopting Gen AI?
The effectiveness of AI to financial institutions hung over the day’s deliberations. This question was asked during the ****opening C-level panel discussion “How to build a data strategy for AI, to mitigate risk and gain business value”, and while the audience answers could have been expected, the panel took a more nuanced view.
Almost two-thirds of respondents said poor data quality was the biggest challenge, ahead of “poor strategic planning” and “lack of explainability, bias and privacy concerns”.
Doris Brophy, managing director and chief data officer for the Americas at Société Générale, echoed the poll results, saying that all AI tools required good quality data inputs, otherwise the outputs would be flawed and potentially damaging.
Chief data officer at Nuveen, Yinghua Michelle Zhou, turned the question on its head, however, asking why AI couldn’t be used to improve data quality. It was a point that BNY chief data officer Eric Hirschhorn had raised in his opening keynote address earlier, explaining that the bank was using the technology to bring structure to unstructured data.
Earlier polls that sought insight into firms’ use of AI found that two-thirds said AI was being actively used and that most value created by AI implementation was being generated in the form of productivity benefits and through product and experience differentiation.
With such benefits already being felt, moderator Julia Bardmesser’s frustration with the data quality issue was apparent: The founder and chief executive of Data4Real bemoaned having to have a long conversation about quality before “I get to have fun with AI”.
At what stage is your firm on the journey to data architecture modernisation?
With new tools and technologies making data management more effective and a source of value, the upgrade of systems has become a matter of great importance to companies wanting to maintain their competitive edge.
The novelty of the concept, however, was illustrated in responses to this poll question, with the largest proportion – almost half of respondents – saying they had started the process of modernisation but only 10 per cent claiming to be in an advanced stage of development.
During the “Designing for the future – how to simplify and modernise data architecture to unleash data value and innovation,” Thomas McHugh, chief executive and co-founder of Finbourne Technology, said his company had seen a lot of customers feeling left in the dark about how to start such a programme.
Part of the problem, McHugh added, was a shortage of understanding of how firms’ legacy systems could be upgraded, a deficit largely caused by the fact that the engineers who had developed and best understood the systems were now retired.
This hunch was backed by findings of a poll held later in the panel discussion, when summit attendees were asked what they found had been the biggest challenges in simplifying and modernising data architecture at their organisations.
While there was a plurality of responses, a third said that “moving away from legacy technology” was their biggest impediment, followed closely by “eliminating data silos”.
Victor Dituro, global investment and risk data lead at JPMorgan Asset Management said there was no getting away from the challenges of modernisation, opining that any legacy transformation would be messy and is “never going to be beautiful”.
Have you established a data marketplace in your organisation?
The same panel posed the poll question that got to the heart of one of the more innovative means of data and data product sharing. As A-Team Group president and chief content officer Andrew Delaney explained, such marketplaces have been in operation for several years and have progressed through the industry.
A third of respondents said they had an internal marketplace but about half, the largest proportion, said they had neither an internal or external structure.
The reason for that, suggested panel moderator Brian Buzzelli, the head of data practice at Meradia, was that there is little understanding of what such marketplaces are.
Nevertheless, marketplaces offer huge benefits to organisations, said Ku at Informatica, which provides such capabilities. When they are structured like an e-commerce platform and offer the same sort of customer experience, marketplaces can democratise data and make its value better understood to the wider enterprise – a point echoed by Mahesh Narayan, institutional asset management segment head at Arcesium.
To what extent has your organisation adopted data products?
With marketplaces providing a repository from which users can access the latest data and data products, the next question followed naturally and was posed during the panel discussion “How to set up a data products strategy to deliver business value”.
While the answer that attracted the single-largest proportion of votes – a third of them – was “to a limited extent”, the rest were almost evenly split between “to a great extent” and “to some extent”.
The need for data products had grown with the number of use cases for, and patterns emerging within, the data, said Leslie Kanthan, chief executive and co-founder of TurinTech AI. Those new tools would enable organisations to derive revenue from their data, Kanthan added.
Brian Greenberg, senior director, business engagement lead for enterprise data management at BNY identified two types of data products – producer-facing tools that ensure data quality, control and other governance imperatives, and consumer-facing offerings that help customers better manage their businesses. Greenberg said he expects the latter to be the next frontier in data services provision.
Subscribe to our newsletter
