Related Videos:
The Patra Perspective, AI in the Insurance Industry - Marcel Gagnon
AI in the insurance industry
The Patra Perspective, Unlocking Efficiency - Brett Cleveringa
Unlocking efficiency
The Patra Perspective, Why Should Agencies Partner With Patra - Sharon Kern
Why agencies partner with Patra
The Patra Perspective, Strengthening Insurance Operational Efficiency - Will Dogan
Revolutionizing the industry
The Patra Perspective, Tony Li discussing Patra's proven method to training accurate AI Models
Patra AI models, insurance evolved
The Patra Perspective, Strategic Program Management for Enterprise Success
Boosting enterprise efficiency
The Patra perspective, case builds for benefits management
Benefits services that deliver
A Message From the CEO, Pratap Sarker
Reshaping the industry

My name is Tony Li, and I’m the CTO here at Patra, and I run the engineering organization.

Insurance workflows are complex, and as good as we’ve gotten to with Gen AI these days, it cannot solve the problem end to end. Gen AI can be very expensive to use in an enterprise level, but they can be be very good at certain things. So we definitely want to make use of it, but not only it.

At Patra, we deal with a lot of insurance documents that are quite complex. So we’re focused on specifically entity extractions. So if you think about a commercial policy with a hundred pages of document, you just want to extract certain key values from it. That is the very time and labor intensive process that the AI can help.

We have thousands of employees who use our AI tools, and we process millions of transactions a year, and we’ve been doing it for [twenty] years. And this is across the retail brokers, the MGAs, and wholesalers. This lets us see into the entire problem space. We have data from a whole variety of documents from insurance companies across lines of businesses.

This is what our competition cannot do. The InsurTechs that are out there, yes, they’re focused using Gen AI, other language models like we are. But what’s key to having these models perform well is to have access to that data. Because they don’t have access to the same data, the quality of the models who just train on that data is gonna be far lower than the ones that we have.

The other thing is also the way that our tools are built: Every piece of work that our people do feeds back into training our models. So both from the completeness of data and just the quantity of processing and retraining that we do makes our models unique and very accurate.

We start our AI journey with policy checking, and we’ve powered it through the AI models we just talked about, both in the full-service model where we take care of everything for you, or self-service model where you log in as a SaaS product. We have learned a lot from our policy checking solution and the AI work that we’ve done can all be generalized. It’s a very exciting time for us right now because we can take these solutions across all different document and workflow types, and you’re gonna be seeing a lot of solutions that matter to you coming out into the market very quickly.

Patra's Method for Training Accurate AI Models in Insurance

Patra's CTO Tony Li reveals how the company's unique approach to AI models is transforming complex insurance workflows.

Patra’s AI innovation enables us to extract key information from complex commercial policies, a traditionally time-intensive process. What sets Patra apart is their continuous learning system—every task completed by their thousands of employees feeds back into training their AI models, creating a cycle of ongoing improvement.

While InsurTechs are exploring similar insurance technology solutions, they lack access to Patra’s comprehensive dataset spanning retail brokers, MGAs, and wholesalers across various lines of business. This data deficit directly impacts model accuracy and performance.

Patra’s journey with AI in insurance began with policy checking, available through both full-service and SaaS models. The company’s AI development framework has now evolved beyond this initial application, with plans to expand these capabilities across different document types and workflows.

Through this strategic approach to artificial intelligence, Patra is positioned to rapidly deliver innovative solutions that address the industry’s most pressing challenges.

Data Advantage

Patra's AI models achieve superior accuracy through access to extensive insurance data collected over 20 years, giving them a significant advantage over InsurTech competitors in developing effective insurance technology solutions.

Continuous Learning

Patra’s continuous improvement approach ensures that every transaction processed by their thousands of employees feeds back into training their AI models, creating a self-reinforcing system that drives ongoing AI innovation in insurance workflows.

Expanding Solutions

After successfully implementing AI in insurance policy checking, Patra is now expanding their AI development capabilities to address additional document types and complex workflows, promising rapid delivery of solutions across the insurance ecosystem.
Find Out More
Tony Li, Chief Technology Officer

Meet our featured expert

Tony Li
Chief Technology Officer

With a commitment to helping Patra expand its focus on technology, software-based processes, and InsurTech solutions, Tony’s strategic oversight and leadership of the technology team will ensure continued innovation and automation of how insurance is sold, managed, and processed. Tony was the Chief of Data Services for PlusAMP and held numerous positions in technology and data services roles.

About Patra

Patra is an outcome-focused operating partner working exclusively for the insurance industry. Combining expert teams, process intelligence, and purpose-built technology, Patra delivers measurable results to clients. With 20 years of deep domain expertise, exceptional client retention rates, and significant sustained investment in AI, Patra partners with agencies, brokers, MGAs, MGUs, and carriers to transform how insurance operations are delivered.