Boosting Alpha

During Third Bridge Partner Day, Jamie Prangnell, Senior Product Manager at Third Bridge, sat down with Mats Stellwall, Principal Architect for AI/ML at Snowflake, to discuss how the Third Bridge <> Snowflake partnership is transforming the way investors leverage expert research through AI.

We’ve distilled the core insights from the webinar to answer key questions for mutual clients. 

Q: What core problem does the Third Bridge and Snowflake partnership solve for investors?

A: The partnership aims to eliminate architectural friction caused by a fragmented data landscape. Currently, many investment teams struggle with data that is stuck in silos or restricted by rigid legacy infrastructure, which bottlenecks actionable research at scale. By integrating Third Bridge’s high-quality expert insights directly into the Snowflake AI Data Cloud, investors can unify diverse, real-time datasets and quickly validate their investment theses without the usual data engineering hurdles.

Q: How can clients leverage Third Bridge insights via Snowflake?

A: Clients can access Third Bridge’s proprietary expert transcript library directly within their Snowflake account. This open approach allows for:

• Instant entitled access: Data is available via private shares, eliminating the need for complex S3 buckets, API calls, or SFTP setups.
• AI-powered analysis: Using Snowflake Cortex AI, users can query unstructured data (like transcripts) and structured data (like SEC filings) using natural language.
• Agentic workflows: Investors can build "Cortex Agents" to automate heavy lifting, such as synthesizing investment themes, comparing company strategies, or even generating investment memos.

Q: How does Snowflake ensure the AI insights are trustworthy?

A: Trust and transparency are built into the platform through several mechanisms:

• Citations: The AI provides direct references to the source material (e.g., a specific broker report or transcript) so users can verify the answer and ensure the model isn't hallucinating.
• Governance and security: The system uses role-based access control, meaning an AI agent can only access data the user is specifically authorized to see.
• Observability: Users can access logs to see exactly how the AI reached a conclusion and provide feedback to improve accuracy.

Q: What makes this partnership different from other AI data solutions?

A: The primary differentiator is Snowflake’s “easy button" approach. Snowflake focuses on removing the technical complexity of setting up and managing large language models (LLMs). This allows research analysts to focus on gaining insights, such as transcribing voice files, extracting data from images, or performing sentiment analysis, without needing to write complex code or manage infrastructure.

Q: How does Snowflake support investment researchers in handling different content types beyond standard text? 

A:  Snowflake provides functionality to process diverse unstructured data formats, including the ability to transcribe voice files from research calls and extract data from images and PDFs. This allows researchers to perform sentiment analysis and combine non-textual information with other research data.

Q: How can investors leverage custom agents as part of their research workflow?

A: Investors can build custom Cortex Agents through a configuration exercise where they define specific tools, such as Python code for web searches or generating investment documents. These agents can be instructed on specific tones and behaviors to automate tasks like synthesizing investment themes or comparing company strategies.

Q: What mechanisms are in place to help people trust the sources of data coming through?

A: Trust is established through citations that link AI-generated answers directly to source materials, such as specific transcripts or broker reports. Additionally, the platform provides observability logs to track the AI's reasoning steps and allows for the use of "LLM as judges" to verify the relevance and accuracy of the output.

Q: Moving data to a data warehouse like Snowflake can be daunting for some investors. How can they overcome this hesitation? 

A: Overcoming this hesitation comes down to two primary drivers: unprecedented accessibility and rigorous security. 

• Accessibility: The big driver for the buy-side is the ability to receive data directly through the Snowflake Marketplace and private sharing. Instead of managing the traditional friction of setting up SFTPs or calling APIs, the data is simply there. As soon as Third Bridge updates a research transcript, it is immediately available for querying within the client's account. The sheer value of this instant, automated access likely outweighs the initial fear of migration.
• Security: Furthermore, Snowflake addresses data safety concerns through its built-in security and governance framework. The platform ensures all data is encrypted and strictly protected, making it inaccessible to anyone outside the specific authorized account. By combining this ease of access with a highly secure, governed environment, the transition becomes less of a technical hurdle and more of a strategic advantage for the firm.

Ready to see how Third Bridge and Snowflake can accelerate your research workflow? Request a demo