Stephen Wood: Rathbones AM COO on data, AI and costs

Stephen Wood took the role of chief operating officer at Rathbones Asset Management in April 2026 after leading its operational and data work since 2021. He has built a Snowflake-based data platform and is running pilot projects with agentic AI to raise front-office productivity and limit cost increases as the firm grows assets under management.

Wood joined Rathbones to design a portfolio management system and then established a formal data strategy to replace spreadsheet and PDF workflows. Over the past year his team has loaded nine datasets into Snowflake covering performance, risk, sustainability and unstructured records of investment decisions. Those datasets feed dashboards and more detailed analytics used by fund managers.

RAM’s technology stack positions Snowflake as the central data cloud. The firm uses Alteryx for ETL and automation, Experian Aperture Data Studio for data quality, catalogue and lineage, MuleSoft for group-level integration and MSCI BarraOne for quantitative risk and performance attribution. Outputs from BarraOne are linked into Snowflake to enable shared data and consistent reporting across teams. RAM retains oversight of its Charles River portfolio management system while outsourcing back-, middle- and trading-office operations.

The asset manager draws on engineering resources from the wider Rathbones group for specialist Snowflake work, while keeping direct access to its own environment and the ability to build local pipelines. Wood reports to RAM CEO Tom Carroll, who reports to Rathbones CEO Jonathan Sorrell, and coordinates longer-term technology direction with group CTO Brad Novak.

RAM is testing agentic software inside its Snowflake environment. The firm has set up a sandbox running millions of rows of data to pilot Snowflake’s CoWork, a knowledge-worker agent, and CoCo, a coding agent for developers and data engineers. The pilots aim to augment constrained engineering capacity and to explore automated coding and knowledge tasks that support portfolio management, product specialists and distribution teams.

Wood groups the AI work with other emerging technologies in a broader strategy that includes quantum computing and tokenisation. He described the approach as thinking about AI, platforms and agents together within that wider plan. He also described his role as not typical for a COO, noting his background spans trading systems, data vendors, investment banks and both front- and back-office functions.

When assets under management rise, many firms add headcount; RAM has prioritised data infrastructure and automation to contain operating cost growth. Wood pointed to the firm’s move away from manual data handling over several years and said the recent work has delivered efficiencies across the business.

The firm’s approach to agentic AI is cautious, with attention to control, compliance and auditability. Wood warned that wider use of agents would change operating models and require subject matter experts to manage and continuously train agents so outputs remain accurate and governed.

Wood’s career includes early work as a data analyst and a role at CQG teaching algorithmic trading, time as global head of market data at Deutsche Bank where he ran a data function, and a period at web technology firm OpenFin before joining Rathbones in 2021. His combined experience in technology and operations has guided RAM’s focus on data foundations and measured trials of new tools.