Revising History: How Vendr Is Conquering Legacy Data

I’m Ken Furie, Director of Data Operations at Vendr, the company that's revolutionizing the game in SaaS contract negotiations. In this intriguing journey, I'll uncover how Vendr overcame the challenges of manual data entry, reshaping our approach to SaaS contracts and unlocking a world of possibilities.

Unearthing the SaaS Goldmine

Vendr has a unique edge in the SaaS landscape. As part of our function as SaaS deal negotiators, we receive SaaS contract data from our diverse range of customers. This data helps Vendr empower customers with the insights needed to negotiate SaaS contracts expertly. But, here's the twist – this treasure trove of data wasn't always as efficient as it is today.

The Manual Maze

For years, Vendr's team navigated a labyrinth of manual data entry. Each SaaS contract was manually transcribed, a process that left room for human error and data gaps. As the number of contracts grew, the situation became increasingly challenging. The manual approach was like navigating a maze with no end in sight.

The Vendr Vision

The Vendr team knew we needed a game-changer. We conceived of a series of ambitious projects designed to overhaul our data operations completely. Our unified goal: an automatic contract extraction system with high accuracy.

The challenges we faced included having to work through 150,000+ documents. The solution had to operate with the highest quality, and it required multiple parallel projects to reach the end goal in a tight timeframe. A key consideration in this process was ensuring that all data was generalized and anonymized along the way, respecting the privacy of all Vendr's customers.

The Transformation Begins

As we embarked on this journey, Vendr was well aware that we weren't just conquering historical data; we were revising history itself by improving the data stored for each deal. The manual errors and data gaps would be a thing of the past.

We designed a trio of interconnected tool solutions, each capable of functioning independently. The first involves optical character recognition (OCR), which feeds its outcomes into the second solution—an AI-based interpretation engine. This engine parses data, identifies specific fields, and structures information to align with our data warehouse schemas. To enhance the interpretation engine, we created coded workflows that handle conditional factors, akin to AI prompting. These solutions are supported by ongoing quality assurance efforts, generating tracking sheets and detailed reports to showcase the extraction's effectiveness and accuracy.

As part of the engine development, we implemented a system capable of recognizing eleven distinct document types, ensuring efficient extraction for each. Vendr adopted an iterative approach for its projects, meticulously versioning extraction software to prevent the introduction of regression errors during code improvements.

We’ll be highlighting more details on our technical extraction solutions in future posts.

A Bright Future for Vendr and SaaS

With our system-wide improvements, and its goal of an automatic extraction capability in the bag, Vendr is now positioned to offer expert insights across even more of the SaaS landscape. We’ve added a high-quality automated self-serve capability to ingest contract data going forward, making it easier than ever for customers to ensure their own data is updated and accurate, resulting in the best possible deal recommendations.

In Conclusion

At Vendr, we've rewritten the script for SaaS contract data operations. We’ve turned a cumbersome, error-prone process into a meticulously managed, data-rich experience. Their journey is a testament to the power of technology, innovation, and a splash of courage to challenge the status quo.