Capital-markets firms are under pressure from growing data volumes, shorter settlement timelines, rising expectations for real-time insight, and AI initiatives moving from pilots into production. The durable advantage is not a model in isolation, it is connecting research, trading, risk, operations, and compliance on reliable, governed data.
In this Q&A, Andrea DeSosa, Global Head of Capital Markets GTM at Databricks, shares a practical approach to modernizing the trade lifecycle around the decisions that matter most.
Key highlights
- Start with high-value decisions: Ask where execution cost is diverging from expectations, how a rate or volatility shock could affect exposures, or which workflows have the highest exception and settlement-fail rates. The questions help define the data foundation needed to answer them.
- Connect the lifecycle’s critical data: Bring together orders, executions, positions, market and reference data, research, risk outputs, client information, settlement status, collateral, funding, and operational exceptions instead of creating another isolated workflow.
- Use AI where governed context matters: AI can help investigate exceptions, analyze execution quality, synthesize research, and surface potential surveillance issues when proprietary data is discoverable, reliable, permissioned, and traceable.
- Make auditability part of the architecture: Production workflows need visibility into data quality, entitlements, lineage, model and agent evaluation, decision records, and the investigation of exceptions, not just a model that produces an answer.
- Modernize incrementally: Firms do not need to replace every system at once. Start with one measurable workflow, establish reusable governed data and controls, prove the outcome, and expand across the lifecycle.
👉 Read the full post here
