TJC’s Transformation: Leveraging Sigma and GenAI to Bridge the Gap Between Unstructured Data and Business Insights

We spoke with Richie Caputo, Director of Data and AI at TJC, a middle-market private equity firm. Richie oversees the firm’s data and AI strategies, both internally and across their portfolio companies. While TJC has always embraced technology, they faced challenges working with large volumes of unstructured data, including complex documents like capital agreements and term sheets. Sigma has played a crucial role in addressing these challenges, enabling TJC’s teams to seamlessly interact with larger datasets, bridge the gap between technical and non-technical users, and leverage AI-powered models to extract structured insights and enable error correction.

Life Before Sigma

At TJC, we’ve always been focused on using technology to make our businesses better, both internally and for our portfolio companies. As the Director of Data and AI, my job is to enable data-driven strategies across the firm. But before Sigma, we were facing a pretty big challenge: our users, who are very comfortable with Excel, were struggling when it came to dealing with large volumes of data. Excel is great for small-scale data tasks, but when you start scaling up, it just doesn’t cut it.

We were constantly bumping up against Excel's limitations, especially with our portfolio companies generating vast amounts of unstructured data—documents like capital agreements, term sheets, credit agreements, and even resumes. These documents are hundreds of pages long and parsing them manually took an enormous amount of time. We needed a better way to extract structured insights from this data without burdening our users with hours of manual work. We tried some other tools but nothing really bridged that gap between the complex data infrastructure we had in place and the ease of use our teams required.

Choosing a BI Solution

What we needed was something that could handle complex data use cases but still be accessible to non-technical users. That’s where Sigma came into play.

When it came to choosing a BI solution, we knew we needed a platform that would integrate seamlessly with Databricks, our existing data warehouse. I’ve been working with Databricks for a long time, so it was important that whatever tool we chose would complement that system. We looked at the usual suspects—Tableau, PowerBI—but none of them really fit the bill for us. What we needed was something that could handle complex data use cases but still be accessible to non-technical users. That’s where Sigma came into play.

Sigma was able to do two critical things for us. First, it provided an interaction layer that made it easier for our non-technical users to visualize and interact with large datasets. And second, it had this unique ability to let users write back directly to the data warehouse, which is crucial when you’re dealing with AI-generated content that requires human oversight for accuracy. The ease of use combined with the robust backend infrastructure Sigma supports made the decision easy. It was the perfect bridge between our complex data environment and the user-friendly experience we needed.

Life With Sigma

After implementing Sigma, it became a game-changer for us at TJC. Our team could now visualize larger datasets in ways that were simply not possible before, and they could do so without ever leaving the Sigma interface. We’re using it to generate structured data from unstructured sources—think of documents like PDFs that contain financial metrics, entities, and date-time data. Sigma makes it easy to extract that information, and our teams can error-correct directly within the platform.

One of the biggest advantages is how Sigma allows us to integrate AI-powered models directly into the workflow.

One of the biggest advantages is how Sigma allows us to integrate AI-powered models directly into the workflow. We’re using large language models (LLMs) to parse these huge documents and extract relevant details, whether it’s from resumes or financial agreements. Sigma is the front-end for all of this—our users can not only visualize the AI-generated outputs but also correct any errors and write that corrected data back into the warehouse. That ability to both read from and write to the data warehouse in one seamless flow has saved us countless hours.

And because we can control permissions programmatically through Sigma’s API, we’ve got complete control over who sees what, and we know our data is secure. The integration with Unity Catalog on the Databricks side makes it even smoother; if someone creates a table in Databricks, those permissions are automatically inherited by Sigma. That’s huge for us from a security standpoint.

What’s Next with Sigma

Looking ahead, we have big plans for Sigma. We’re exploring more advanced AI-driven use cases where we can further automate the extraction and structuring of data from unstructured sources. For example, we’re working on refining how we process resumes, allowing us to quickly filter out top candidates based on key metrics, while preserving privacy by automatically stripping out personal identification.

The next big step is to fine-tune these models even further, so they can handle increasingly complex data sources. We’re also looking at how Sigma can help us expand the use of AI across our portfolio companies, bringing the same level of efficiency and ease-of-use to their workflows.

Sigma has transformed how we work with data at TJC.

Sigma has transformed how we work with data at TJC. It has allowed us to focus on what really matters—making smart, data-driven decisions—while eliminating much of the manual, time-consuming labor that used to bog us down. And as we continue to build on these capabilities, we’re confident that Sigma will play a central role in helping us scale our data strategy across the firm.