Data management provider GoldenSource has entered the next phase of its evolution with the launch of GoldenSource Scout, an AI platform for capital markets built on the firm’s Trusted Contextual Data Layer, which grounds enterprise AI in governed, canonical data. The launch marks a significant shift towards AI within the firm’s data management ecosystem.
Swati Tyagi, Chief Product Officer at GoldenSource, said the industry has been grappling with the challenge of implementing AI within deterministic operations. While many firms have secured budgets for AI, they have often struggled to apply it effectively to core functions like pricing, risk and reference data.“Everybody will tell you, ‘Yes, we have a fancy AI tool,’ but it will only give you answers based on how good your data is,” Tyagi told Data Management Insight. “We’re not assembling context when a question gets asked. It’s already there. Governed, mastered, client-specific, built into their EDM. Scout makes that usable by AI. Most vendors don’t have access to that kind of context. That’s the trust gap, and Scout closes it.”
Governance Gap
GoldenSource Scout runs on the firm’s Trusted Contextual Data Layer, which connects governed data with its lineage, relationships and business meaning across the company’s data management platform and its Snowflake-native analytical application, GoldenSource OMNI.
Using NLP and generative AI, the platform lets business users ask questions across data domains in plain English, trace values back to source and identify which positions, processes or reports an answer is derived from.
By integrating AI directly into the data layer, GoldenSource aims to solve a long-standing industry tension between chief data officers trying to enforce governance and operations teams burdened by manual, repetitive tasks.
With GoldenSource Scout, the goal is to shift the focus from data maintenance to intelligent decision-making, allowing business users to interact with their data ecosystem more intuitively than ever before.
Auditability and Explainability
The release follows GoldenSource’s recent InvestOps 2026 research, entitled “The Basis Point Blind Spot”. which found 98% of firms are concerned that weak data infrastructure could lead to incorrect AI insights, missed opportunities or financial losses.
Scout is designed to provide more than just data retrieval; it adds context. For example, if a user observes that an ISIN has dropped out of their active security master, Scout can trace the lineage of that record to identify the corporate action, such as a merger or a maturity, that retired it. This capability offers the auditability and explainability that Tyagi said are essential for operational trust.
“Operations shouldn’t be spending time chasing down what happened to an ISIN by writing database queries or raising vendor tickets,” Tyagi said. “Scout tells you why an exception occurred and how to prevent it. It brings the context, the auditability, the explainability, grounded in the client’s own data.
Assigned Agents
Scout is delivered as a set of purpose-built agents, each engineered for a specific workflow. The first release includes agents for data investigation, exception management and lineage tracing. These are already in use inside GoldenSource itself, accelerating support and service delivery, acting as what Tyagi calls the firm’s “most intelligent employee”.
GoldenSource plans to expand Scout significantly across future releases. Capabilities on the roadmap include:
- Unstructured Data Pipelines: An agent capable of reading, deciphering and mapping incoming documents, such as capital calls, directly into the GoldenSource data model and data platform.
- Ratings Mismatch Management: A tool to analyse logic conflicts between different rating agencies, helping users understand why certain ratings are applied and how they align with institutional policies.
- ID Resolution: Advanced matching agents to better organise counterparty and client data that lack standard identifiers.
Phased Growth
The company has an early adopter programme underway with around 10 early adopter clients. As the system matures, the vision is for GoldenSource Scout to become the primary way clients consume and interact with their governed data, with the Enterprise Data Management/EDM platform running underneath.
“Our goal is to make Scout the only way you access EDM over time… so it starts with being an assistant, and then it becomes the way you master data,” Tyagi said.
Subscribe to our newsletter
