Secure RAG platform with chat services powers faster, smarter investment decisions for firm’s clients
Bay Point Advisors, an Atlanta-based investment management firm, specializes in private-market opportunities often overlooked by traditional investors. In this space, speed matters, but researching opportunities is anything but fast. Secure data rooms hold hundreds of PDFs, spreadsheets, and contracts, creating a mountain of information that advisors and clients must sift through before making critical decisions. This manual process slows deal flow and risks missed opportunities. Bay Point Advisors needed a way to accelerate decision-making without sacrificing accuracy or security. Their goal: build a platform that lets clients interact conversationally with data rooms and receive grounded, citation-backed answers in milliseconds.
Hours to minutes
speeds investment analysis capabilities
Minutes to seconds
yields ultra‑fast customer response times
Self-service empowers
enhancing customer experience and satisfaction
To overcome delays caused by manual document review, Bay Point Advisors partnered with IBM for expertise and guidance. “They didn’t just answer questions—they opened a Slack channel, scheduled weekly meetings, and even helped us build a proof of concept,” said Rick Atkinson, CTO. With IBM’s support, the firm-built Atlas, an AI-powered platform enabling conversational access to secure data rooms. Leveraging IBM® DataStax® Astra DB within IBM watsonx.data®, Azure Blob Storage, Azure OpenAI, and a SQL metadata layer, Atlas uses a Retrieval-Augmented Generation (RAG) architecture to transform static documents into a searchable knowledge layer. This integration automated workflows, replacing time-consuming manual review with instant, citation-backed answers delivered in milliseconds. IBM’s direct support accelerated development, guiding retrieval workflows and embedding parameter tuning to ensure security, scalability, and performance. “We couldn’t have completed this project without IBM’s Astra DB team—their hands-on support and guidance made the difference, accelerating development and ensuring success,” said Rick Atkinson.
With Atlas, Bay Point Advisors reduced document review and research time from hours to seconds, accelerating due diligence and improving client responsiveness. The platform delivers grounded, citation-backed answers that cut error rates and minimize hallucinations, ensuring higher confidence in investment decisions. Advisors report significant productivity gains as manual scanning is replaced by automated, conversational queries. Governance remains strong with document-level permissions enforced end-to-end, protecting sensitive data. Astra DB within watsonx.data provided the secure, scalable foundation for this transformation, while IBM experts guided workflow design and parameter tuning to optimize performance. “We couldn’t have completed this project without IBM’s Astra DB team—their hands-on support and guidance made the difference, accelerating development and ensuring success.” Looking ahead, Bay Point Advisors plans to integrate agentic AI capabilities for proactive intelligence. Atkinson remarks, “Atlas delivers a secure, personalized experience for every client. By combining structured and unstructured data with AI, advisors can provide tailored guidance while protecting sensitive information—a true win-win for trust and performance.”
Bay Point Advisors is an Atlanta-based investment firm founded in 2012, specializing in creative investment solutions primarily in the private debt market across the United States. The firm focuses on private-market opportunities often overlooked by traditional investors. Bay Point Advisors aims to accelerate clients’ ability to make critical financial decisions by leveraging advanced technology.
Legal
© Copyright IBM Corporation. January, 2026.
IBM, the IBM logo, DataStax and watsonx.data are trademarks of IBM Corp., registered in many jurisdictions worldwide.
Microsoft, Windows, Windows NT, and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries, or both.
Examples presented as illustrative only. Actual results will vary based on client configurations and conditions and, therefore, generally expected results cannot be provided.