SirionLabs establishes US foothold to scale its NLP contract management software

SirionLabs, a startup providing vendor management software to enterprises, is adding a U.S. headquarters to its footprint. The company was initially founded in India and has raised $16.95 million from Sequoia Capital India, Canopy Ventures and Qualgro VC to extract data from contracts to ensure transparency and accountability.

The establishment of a U.S. presence represents a strategic shift in the company’s growth plans. Last year, SirionLabs told TechCrunch that it was looking to hire and relocate workers from Silicon Valley to India. While the startup has had offices in the U.K., Germany, Denmark and Singapore, it has been slow to establish a permanent U.S. team. The new headquarters will be located in Dublin, California.

Sirion, the company’s platform, is currently used by companies like BP and Vestas to manage service providers and augment humans that traditionally manage vendor relationships. The startup expects to use natural language processing to analyze more than $8 billion in total contract value over the next year.

Ajay Agrawal, CEO of SirionLabs, told TechCrunch that 2-3 percent of a given contracts value is spent on managing the supplier relationship. After a stint as an EIR at Stanford, Agrawal decided to dedicate time and resources to partially automating this process.

“Given the diversity of invoicing practices, we had to come up with a form to detect discrepancies,” said Agrawal. “The contract tells you what should happen and performance data tells you what should have happened.”

The idea of using NLP to analyze and diligence contracts isn’t new. SirionLabs was founded back in 2012, and competitors like Seal Software date back even earlier, to 2010. The long tail of potential discrepancies that need to be flagged has been a bit of a barrier to adoption in a market that doesn’t have a stomach for risk. This is why Sirion requires both sides to agree to common outcomes in person, placing final accountability on humans rather than machines.