Scaling AI in financial services starts with getting the data right

Although 63% of financial services firms have at least a level three maturity for “responsible AI” for data and technology according to a recent survey by McKinsey, more progress is needed. To achieve the data governance and agility that’s required to support AI use cases that truly deliver value, organisations need to rethink their data infrastructure. LSEG understands what is required of this data transformation programme because it has recently undertaken a similar journey, to enable the industry to access AI-ready financial data. 

• Delivering value through an AI strategy requires data that is high quality and accessible. 
• Transforming legacy data infrastructure can be challenging, but LSEG has already undertaken this, to make its market data available for today’s AI use cases. 
• To implement its own AI strategy, LSEG partnered with Microsoft. Today its market data is available within an ecosystem of solutions designed to support AI use cases from ideation to delivery. 

At Microsoft AI Tour London, industry leaders discussed the fundamental role that data infrastructure plays in a successful AI strategy. Accurate AI outcomes is underpinned by data that is high quality, accessible, permissioned, and sits within a data governance framework.

This data must sit within a robust ecosystem that converts information embedded across workflows and analytics. For many firms in the financial services industry, achieving this is full of complexity, thanks to legacy data and technology stacks that have grown organically over time.  LSEG appreciates the challenge financial firms face in transforming their data for AI, because it has already taken this journey itself, by deploying an ecosystem of Microsoft solutions. 

According to a recent survey by McKinsey[note1], 63% of financial services firms have at least a level three maturity for “responsible AI” for data and technology, compared with 55% of all industries included in the study. Responsible AI includes a significant focus on data quality, because AI outcomes are only as good as the data that they are grounded on. It is positive news that the industry has made such progress. 

However, more work must be done to deliver value from AI strategies. Poor quality data will deliver inaccurate AI outcomes, resulting in increased financial, compliance and operational risks – as well as the need for additional human intervention. To build a culture that embraces AI and achieves its strategic goals, a data governance programme that ensures data accuracy and provides the right semantics is needed. Robust data governance helps to build a culture of trust in the ability of AI to automate processes and provide analytical insights.

In addition to data governance, financial firms require agile data to support AI. Most organisations are sitting on massive quantities of their own data, locked in silos. To be able to transform AI into value, firms need to be able to translate this data into intelligence that can flow through processes, deliver insights, generate operational responsiveness, and support regulatory confidence. 

Unfortunately, most financial firms have not achieved this yet. According to another survey, 83% of senior business leaders said their organisation’s AI adoption would be faster if they had stronger data infrastructure in place. [note2]

The complexity of legacy data stacks and associated technology often means there are data copies and infrastructure inefficiencies, limited interoperability between data sets and across platforms, data security challenges and vendor sprawl. 

The solution is to build a single, organisation-wide data lake that enables AI adoption beyond specialist teams, drives productivity and use‑case innovation, and supports enterprise‑wide impact. This is because having a single source of truth for data across the enterprise ensures that everyone is working with the highest quality, freshest data with accurate permissions and essential metadata baked in.

“When organisations move away from having segregated data sets sitting in garden sheds and under floorboards, and instead in a single location that everyone can access, the capacity for AI to transform the business grows exponentially,” said Emily Prince, Group Head of Enterprise AI, LSEG, speaking at the Microsoft AI London event. “When we did this at LSEG, we started to have a view on information that was incredibly powerful in terms of the insights that it could yield.”

By creating a data lake, people can deploy data into AI use cases knowing that quality is strong, and they have permission to do so. People can also access a wider range of data sets, potentially uncovering news, reference data, pricing data, and more that will enhance the accuracy of AI outcomes. Adding in more historical data – which captures periods of financial distress – can improve the robustness of stress tests and generate improved scenario analysis. 

This unified data infrastructure delivers smarter decisions and scalability. End-to-end transparency and governance reduce risk and maintain data trust. While simplifying management and reducing costs, unified data unlocks the ability for employees across the business to use AI to innovate to improve productivity and generate insight.

LSEG’s AI strategy focused on delivering trusted, licensed, AI-ready content to scale AI in financial services. The unparalleled depth, breadth, and quality of LSEG’s AI-ready content and taxonomies includes proprietary datasets stretching back over decades. LSEG Everywhere includes deployment of the Model Context Protocol (MCP) and recent partnerships with Microsoft, Claude, ChatGPT, Snowflake and Databricks.  

“Bringing data to people in a turn-key way, that enables them to ideate and experiment, is extremely powerful,” says Prince. “Now, put that together with Microsoft and MCP, and firms can now work with over 33 petabytes of trusted data that they can really lean into.”

LSEG data sits within the Microsoft ecosystem of data and AI solutions, including Microsoft Foundry (AI), Microsoft Defender (Security), Microsoft Purview (governance), and OneLake. Together, these solutions support data quality and governance. Data rights are embedded, so that end users know what their permissions are. Data discovery becomes more intuitive, enabling AI innovation.

With this infrastructure, AI value can truly be delivered. “We are co-developing with customers, and some of the things that we are able to see, and build are so exciting” says Prince. “Historically, there are so many possibilities that the financial services industry hasn’t modelled, hasn’t captured. Put another way, there is so much opportunity.”