By Byron Fernandez
TDCX Group CIO and EVP

Insurance isn’t merely a financial product, but a promise of protection often made in high-stakes situations. However, a McKinsey study found that nearly two-thirds of fraud-related transaction declines were actually legitimate and led to unnecessary disruptions. Overly rigid fraud controls can reduce fraud losses, but they also create unnecessary friction and frustration for policyholders.

Customers rely on their policies when it matters most, which is why improving customer experience (CX) has become the top feature for enterprises investing in know-your-customer (KYC) and identity verification solutions. To break out of this paradox, insurers need an “intelligent,” more adaptive strategy that entails proactive fraud detection and prevention optimized by modern technology and enabled by human frontliners.

Strategies and best practices for balancing CX and fraud prevention

A more adaptive, intelligent approach involves moving from a trade-off mindset to a holistic approach where the balance between fraud prevention and CX is supported by AI and analytics:

Implement adaptive, risk-based segmentation. This approach streamlines CX for low-risk customers while adding scrutiny for high-risk cases. For instance, AI-powered authentication models analyze customer behavior, past claims, and transaction history to assess risk dynamically. This is critical as 60% of decision-makers cite reducing friction in authentication as their biggest challenge. A study on a healthcare insurance platform using risk-based models also reported the benefit of being able to personalize products and services for customers.

Embed fraud detection throughout the customer journey. Insurers started integrating monitoring at multiple touchpoints, from onboarding to renewals, to better detect patterns before they escalate. This aligns with shifting consumer expectations where nearly 70% of consumers in the Americas place a higher value on a smooth onboarding experience.

Shift reactive response to predictive prevention. Predictive analytics models can flag potential fraud before a claim is even filed. In the US, 74% of insurers have invested in predictive fraud modeling to fast-track genuine claims while filtering out fraud attempts before payouts occur.

Establish transparent, customer-focused communication. Fraud prevention measures can erode trust if policyholders don’t understand why they’re happening. Providing proactive alerts, updates, and self-service options create a smoother experience while reinforcing transparency. This gap is a challenge, particularly for 44% of insurers undergoing digital transformation, admitting their lack of skilled fraud and CX experts. 

Adopt an omnichannel strategy. Insurers need a strategy that can be executed seamlessly across digital, call center, and even in-person interactions, as fraudsters could exploit gaps between different interaction channels. For example, more than 50% of automotive insurance customers in the US prefer digital-only interactions for simple claims but would opt for a hybrid digital-human approach for complex cases. In Asia Pacific (APAC), nearly 90% of insurance customers across generations rely on mobile devices to manage policies. Millennials and Gen Z customers are more engaged in social commerce, using social media platforms for shopping and researching insurance options.

Continuously update fraud detection models. A model that works today might be obsolete in three months, particularly as AI-generated fraud, synthetic identities, and deepfake claims become more prevalent. To illustrate: Nearly three in five businesses already using AI and GenAI retrain their models quarterly to keep pace with evolving large language models. Some commercial AI models even get released or updated monthly.

Train and upskill employees. A well-trained team reduces false positives, improves fraud detection accuracy, and ensures smoother interactions and experiences. In a case study, an insurer trained 35% of its employees and nearly all its agents in CX-focused fraud prevention, resulting in a 25% reduction in processing times, improved cross-sell rates, and increased customer retention.

Here’s how AI and analytics enable these strategies:

Automated risk scoring: AI can streamline the process of analyzing how customers interact with digital platforms, ensuring that low-risk interactions remain frictionless while high-risk cases are flagged for further verification. This precision matters, as 51% of consumers in the US say that they would abandon an insurer if the claims process felt unnecessarily complex.

Machine learning-powered (ML) pattern recognition: ML models continuously improve detection accuracy by recognizing subtle patterns across claims and policies. In fact, a company demonstrated how using ML helps uncover three times more fraud than manual or rules-based approaches. Additionally, 90% of financial services leaders reported that their company's use of ML models resulted in higher confidence in their detection accuracy.

Data-driven learning and training: When employees have access to data-powered tools, they can identify red flags faster and make informed decisions. A case study showed that when a government agency integrated data and analytics into their insurance disbursement workflows, it increased fraud detection by 60% while simultaneously reducing false positives by 50%.

Framework for balancing technology, data, and human expertise

No single factor is sufficient on its own. Overreliance on one creates blind spots: Automated systems can miss context, manual reviews can be too slow, and fragmented data can undermine accuracy.

Technology enables speed and scalability. In fact, insurers using AI-optimized claims processing already report up to a 73% increase in cost efficiency, thanks to AI’s capability to analyze vast amounts of transaction patterns and behavioral data. However, fraudsters can adapt to detection methods as quickly as they’re introduced, which makes human expertise indispensable. 

Human expertise provides depth and context. While AI models flag potential fraud, fraud analysts and CX specialists assess edge cases and refine detection models. One of the leading ride-hailing companies in North America, for example, utilizes GenAI models alongside their CX frontliners to provide customer service, with support for user safety, account deactivations, and fraud still analyzed and managed by human agents. 

Their insights can also feed back into AI systems, improving accuracy and reducing false positives. This is what a multinational financial services firm has been doing for years, helping its AI tool learn from inaccuracies and human reviews/audits to better identify risk factors.

Data is the foundation that powers AI models and informs human decision-making. Properly utilizing it remains a perennial challenge, however. Nearly 90% of financial services leaders acknowledge that advancements in technology have made it easier to collect customer feedback and behavioral data. Yet, they also admit to not knowing how to properly organize, analyze, and act on this information. This is exacerbated by how 60% of financial executives struggle to find the right tech stack to process their data.

Effective fraud detection and prevention requires a cohesive strategy that balances technology, data, and human expertise. It’s this framework that enabled a global tech company to scale its fraud detection and customer support operations across more than 13 markets in Southeast Asia, South Asia, and East Asia. A localized approach — combining multilingual expertise, market-specific training, and AI-assisted workflows — helped the company surpass industry standards, achieving over 90% in quality performance while continuously refining their own monitoring guidelines through focus groups organized monthly.

Achieving this balance is a competitive advantage. An intelligent approach will not only strengthen defenses against fraud but also build long-term policyholder trust and retention.