Product descriptions:
RBC BD helps people across the UK and Ireland plan and shape their financial futures by protecting, growing and managing their money. We believe in the value of ongoing financial advice to help our clients achieve their goals. An annual meeting is an excellent opportunity to review clients' circumstances to ensure we provide a service that best suits their needs. To streamline preparation for a client review meeting, RBC BD turned to the Databricks Data Intelligence Platform to help unify data across CRM, trading systems and advisor notes, while automating the information gathering process. By integrating GenAI for document parsing, insight extraction and dynamic report generation, RBC BD seeks to create office capacity (an estimated savings of 4,700 hours annually) and simplify the information gathering process.
Eliminating guesswork to standardize client reviews
RBC Brewin Dolphin, a leading wealth management firm, serves a wide range of clients—from individuals and charities to corporations. With rising client expectations and increased industry disruption, the firm sought to modernize the process for preparing Client Advice Reviews (CARs), which plays a critical role in the client’s annual financial planning. These reviews required a complete picture of each client’s financial life—from portfolio performance to risk profiles—but preparing an advisor for the annual client meeting and assembling them was a manual, time-intensive process. Data lived across five disparate systems, including CRM tools, trading platforms, SharePoint, and locally stored advisor documents. There was no consistent format or method for compiling data for this annual client meeting, as some used Excel, Word, or simply relied on memory.
The lack of standardization not only consumed time but also eroded advisor trust. “Sometimes advisors would spot a number in the report and think, ‘That doesn’t look right,’ but there was no easy way to figure out where it came from,” explained Dr. David Elliott, Lead Data Scientist at RBC Brewin Dolphin. Without transparency or traceability, it became harder and more time-consuming to validate data.
Compounded across 14,000 annual reviews, this challenge pulled advisors away from high-value client conversations and into administrative work. To automate content generation, apply GenAI, and improve overall efficiency, the firm first needed to unify its data and modernize the underlying infrastructure.
Modernizing preparation for a client's annual review meeting
To modernize the CAR pre-meeting preparation process, RBC Brewin Dolphin first needed to address the fragmented nature of their data. With client information dispersed across numerous systems, the firm turned to the Databricks Data Intelligence Platform. Using Delta Lake, RBC Brewin Dolphin centralized both structured (e.g., portfolio and risk data) and unstructured (e.g., meeting notes, uploaded PDFs) data. The bronze, silver and gold layers allowed them to ingest raw inputs, clean and join data and produce business-ready outputs for analytics and AI.
Unity Catalog provides row-level security and fine-grained access controls. Integrated with the firm’s enterprise SSO system, Unity Catalog ensures that advisors only access the data relevant to their own clients, helping the company meet its rigorous compliance and privacy requirements in a highly regulated industry. On the application side, RBC Brewin Dolphin built a custom Databricks App using Dash, a Python framework for data apps. The interface lets advisors upload documents, trigger workflows, and more easily prepare documents for client account review meetings from a single, user-friendly dashboard. Crucially, the app mimics familiar tools like Word and Excel, reducing friction and enabling faster adoption among internal advisors.
Behind the scenes, Lakeflow Jobs orchestrates a series of GenAI pipelines that automate the most labor-intensive parts of the CAR process. Each session runs 30 to 40 tailored LLM prompts, primarily using Llama 3.1/3.3, optimized for document type and content. This allows the system to parse context-rich narratives and generate insights advisors can review before meeting with clients. The team also integrated Vector Search and retrieval-augmented generation (RAG) to support more contextual, advisor-ready outputs. As advisors upload documents, temporary vector indexes are created on the fly, enabling models to organize information into discussion-focused summaries.
“Our workflows run over ten models in parallel, facilitating rapid processing and discovery of insights that set our financial advisors up for a successful meeting,” said Elliott. “Every number, statement and recommendation to financial advisors is traceable back to its original source, whether it's a CRM field or a line in a PDF. That transparency is game-changing for auditability and advisor confidence.”
Delivering reviews quicker, more efficiently and at a lower cost
By automating approximately 90% of meeting pack preparation, Financial Planners and their support teams can quickly gather the information needed for a client’s annual review. On average, Databricks automation saves 20 minutes per meeting pack, adding up to 4,700 hours saved annually across all packs generated. Delivering fast and comprehensive meeting pack outputs in-house is also expected to reduce administrative costs by 50% for client reviews across our 33 offices.
With a strong data foundation in place, RBC BD is expanding its use of Databricks to explore natural-language interfaces, integrate real-time market data, and build next-generation tools that could help Wealth Managers deliver tailored recommendations more efficiently. For example, the financial services company hopes its investment analysts can fine-tune embedding models, setting the stage for even more relevant outputs personalized for each client.
“We’re starting to look at how we can bring portfolio data to life with context that actually matters to each individual client that invests in our services,” Elliott concluded. “This isn’t just about operational efficiency—it’s about making the experience better, more human and more impactful for our clients.”