A Deeper Look at Sirion’s AI Auto Extraction Capabilities
- Last Updated: Dec 17, 2025
- 15 min read
- Sirion
With every iteration and improvement, AI has been consistently challenging and changing how we live and play – from Alexa taking dinner orders to Google’s DeepMind beating the World Go Champion. With SirionAE however, we wanted to use the power of AI to transform how we work.
Eight years ago when we established Sirion, our singular goal was to develop a SaaS-based platform that would transform how businesses manage contracts and relationships. As a result, synthesizing our own machine intelligence solution had been a central part of our roadmap.
The earliest iteration of our solution was a regular, rule-based engine that the engineering team built to ‘mechanically’ extract legal provisions from contract documents. This proved to be a challenge later because contracts are inherently chaotic, unstructured data sources. These documents are often scattered in siloed systems, email, and sometimes even exist as handwritten notes. Over time, there can be hundreds of amendments, SOWs and work orders tagged to one MSA or framework agreement.
Our conclusion? A rule-based engine wasn’t going to be a long-term solution for how we had envisioned the product.
Building the Next Generation of CLM Tech
Our focus firmly remained on getting the basics right when we were developing machine learning models for the text extraction engine, which has been through scores of iterations over the past several years. We synthesized the engine using a variety of Bayesian network, named entity recognition and n-gram language models along with syntactic rules. Soon enough, we built prototypes and started testing some of the earlier AI-driven features in collaboration with select customers.
Fast forward to 2019-20, and we have now filed several patents and rolled out a stand-alone AI solution for contract management that helps businesses gain full visibility into and control of risks and obligations. What started out as a simple idea to automate document extraction has evolved into something much more. Today, our customers can use SirionAE to process up to a million documents a day. We have transformed our AI engine from a backend efficiency enabler into a customer-facing, self-service solution. Some of our clients are already training Sirion’s AI engine themselves to extract specific fields and clauses, which means that the solution keeps evolving well after it has been deployed, learning about a new structure or clause language with every contract it ingests.
From Document Extraction to an Extraction Review Platform
Somewhere along the way, we had realized that our AI engine’s potential would be underutilized if we restricted ourselves to data extraction and stopped short of delivering an end-to-end solution by bookending our extraction capabilities with data preparation and review features. As it currently stands, SirionAE has matured into a powerful, omni-point solution that fulfils our customers’ AI needs without having to navigate the complexity of using multiple disjointed systems. And I’m also happy to announce that the numbers are in, and our customers are reporting as much as a 50% uptick in the tasks they have been able to automate with Sirion.
Our technology roadmap for SirionAE continues to be ambitious, and we are making steady progress on real-time capabilities designed to deliver powerful insights and enable better business decisioning at the right time, every time.
I am humbled by the incredible work by our engineers, and I look forward to crossing many more product milestones with them as we embark on the next phase of our journey. This is just the beginning and I can’t wait for what’s ahead.
To learn more about Sirion’s auto-extraction engine visit the Sirion product page.
Additional Resources
What is Contract Data Extraction: 5 Key Steps to Implement it
