How Techcom Life deploys AI as only 15% of use cases scale

Its in-house assistant Tori is already being used by about 85% of 1,500 agents.

Vietnam-based Techcom Life Insurance Company Limited is building an artificial intelligence (AI)-first operating model to move beyond pilots and into core insurance functions amidst an industry-wide execution gap where fewer than 10% to 15% of AI implementations are being used at scale.

Ravindra Venisetti, chief information officer of Techcom Life, told the audience at the Asian Banking & Finance and Insurance Asia Summit today in Ho Chi Minh City, Vietnam, that the broader insurance sector is still struggling to convert experimentation into enterprise deployment, even as investment in AI accelerates. 

The rapid adoption of AI in Southeast Asia’s insurance sector is shifting competition toward speed, efficiency, and data-driven underwriting, with new entrants such as Techcom Life positioning themselves as “AI-native” players to challenge incumbents. 

Venisetti outlined how the firm is embedding AI across its operations just three months after launch.

“The plan for Techcom Life is basically not to increase the number of people to thousands of people. So we intend to use technology, AI to run our business. Our objective is to have an AI-first, AI-native organisation,” he said. 

The approach is notable in a market where insurance penetration remains around 2%, significantly below mature markets, creating room for technology-led expansion. 

“There is a big opportunity for Vietnam to do a leapfrog in the AI space because we don’t have to go through the legacy channel and the maturity model that mature countries have gone,” Venisetti said.

Industry data supports this view. According to a 2025 report by McKinsey, AI-driven underwriting and claims automation could reduce operating costs for insurers by up to 40% while improving loss ratios by three to five percentage points. 

Meanwhile, Deloitte estimates that global insurance AI investments surpassed $10b in 2024, with Asia-Pacific accounting for a growing share as digital-first insurers scale.

However, Venisetti cautioned that industry-wide implementation remains limited. 

“We are talking about less than 10% to 15% of implementations [being] used at scale,” he said, highlighting the gap between experimentation and enterprise deployment. 
Legacy systems remain a key constraint in developed markets.

“Some of these companies have been running for 100 years. It’s almost very impossible for architecture to scale up to run your AI because your data is all over the place,” Venisetti added. 

In contrast, Southeast Asian insurers benefit from cloud-native systems and simpler architectures, enabling faster integration of AI tools.

Techcom Life is focusing on core insurance functions such as pricing, underwriting, and claims.

“The core of the business for insurance is the pricing. For hundreds of years. This has been pretty much on paper Excel sheets. I think AI is going to change this,” Venisetti said.  

The company is also deploying AI in customer acquisition through real-time sales support tools. 

Its in-house AI assistant, “Tori,” is already used by approximately 85% of its 1,500 agents, providing live prompts during client interactions.

Fraud detection is another priority, particularly in Southeast Asia, where fraudulent claims are estimated at 10% to 20%. 

Whilst AI cannot eliminate fraud entirely, Venisetti said it enhances pattern recognition and early detection capabilities.

Operationally, the company is testing extreme automation. Venisetti revealed that an AI-generated document management platform was built and deployed “in three days” with “zero coding by humans.”

He added that AI agents are already functioning as virtual employees, including administrative roles and software development collaboration.

Despite the momentum, internal prioritisation remains a challenge. 

“We have way too many ideas. every time. We come up with 10 to 12 new ideas. And we just don’t know which one to actually start with,” Venisetti said, identifying execution sequencing—not technology—as the main constraint.