Retirement service providers are under increasing pressure to deliver personalized, secure, and scalable support. With growing participant bases, rising expectations, and a complex financial ecosystem, customer service in the retirement sector is evolving. A key driver of this transformation is the integration of Large Language Models (LLMs) and Small Language Models (SLMs), which are reshaping how contact centers operate. By automating routine queries, assisting human agents, and improving turnaround times, these AI models are helping providers meet demand while maintaining trust and compliance.
This blog explores how AI is changing the dynamics of 401(k) contact centers and why combining digital intelligence with human empathy is the way forward.
smarter bots, smarter workflows
Traditionally, participant queries were addressed by either service representatives or rudimentary bots. While human agents brought empathy and critical thinking, early bots were limited to predefined scripts and offered minimal support. The integration of LLMs and SLMs has significantly expanded the capabilities of bots, enabling contextual, accurate, and real-time responses.
Today’s AI-enabled bots can:
- Resolve routine queries and provide 24/7 assistance
- Summarize conversations and pre-fill workflow forms
- Offer real-time suggestions to human agents
These capabilities not only improve first call resolution (FCR) but also reduce average handling time (AHT). Yet, challenges remain. Live access to financial data raises data privacy and security concerns. While SLMs offer localized control, they often lack the comprehensive accuracy of LLMs. On the other hand, deploying LLMs at scale requires significant investment in infrastructure, compute power, and specialized talent.
Moreover, not all queries can or should be handled by bots. Complex, emotional, or sensitive issues still require human support. Manual processes such as note-taking and ticket routing, however, introduce inefficiencies. AI can mitigate this by assisting with self-service virtual assistants and chat/voice bots, language and ascent neutralization, automated call summaries, structured data generation, predictive & intelligent routing, persona dashboards and much more.
When combined with NLP, real-time transcriptions can automatically capture call details, providing an automated call summary that covers all essential call details and ensures the reliability of customer records. This minimizes manual note taking for agents, reducing the risk of errors, enabling quick action by the agents, and improving resolution efficiency in call centers.
rising demand in the 401(k) ecosystem
The U.S. Defined Contribution (DC) system underscores the scale and complexity of retirement services. As per the 2024 Recordkeeping Survey by PlanSponsor:
- $10.9 trillion in assets under management as of December 2023
- 129.6 million participants
As per the ‘The US retirement industry at a crossroads’ article by McKinsey:
- $39 billion in annual revenue, up from $28 billion in 2013
- 11.1 million Americans aged 65+ projected to remain employed by 2025
As the participant base expands, so do their service expectations. In Q2 2024, Shwab Survey finds that:
- 61% of participants expressed a need for professional financial advice, up from 55% the previous year
- 60% reported feeling overwhelmed by plan information
- 77% desired more support in making informed decisions
While digital tools are increasingly common, many users, especially older participants, still prefer human assistance when navigating complex retirement decisions. This points to a need for a hybrid model: intelligent automation combined with empathetic human service.
managing call volumes and service expectations
Call centers in the retirement space face seasonal demand spikes. For example, as per the recent study of a global investment management firm, there is a constant surge in calls each January due to contributions and withdrawals. The United States Social Security Administration (SSA) manages nearly 400,000 calls per day, with waiting times exceeding 90 minutes and a 46.1% answer rate (as published by Kiplinger).
High abandonment rates in call centers, often between 5% and 10%, are primarily driven by:
- Long hold times
- Inefficient IVR design
- Multiple call transfers
Most participants are unwilling to wait more than five minutes, emphasizing the need for streamlined service workflows. To remain competitive, providers must proactively invest in AI, improve call routing, and optimize workforce scheduling.
the expanding role of chatbots
The global chatbot market is projected to grow from $15.57 billion in 2024 to $46.64 billion by 2029 (as per the Chatbot Statistics 2025 report by Exploding Topics). According to Gartner, 80% of organizations plan to use AI-powered chatbots for customer service by 2025. Within the U.S. retirement services segment:
- 66% of recordkeepers currently use AI solutions
- 33% are planning future implementations
Well-implemented chatbots can deflect 30% to 60% of incoming queries. However, the real strength lies in seamless escalation. Bots must recognize when to hand over to a human agent, complete with full context and conversation history. This minimizes frustration and ensures continuity.
security, trust, and ethical AI
With the rise of AI, the risk of scams and data misuse increases. Given the high value of retirement accounts, maintaining trust is non-negotiable. Providers must:
- Implement strong data encryption and access controls
- Use explainable AI models to ensure accountability
- Maintain human oversight in sensitive interactions
- Conduct regular security audits and compliance checks
Balancing automation with ethical and secure practices is crucial. AI must support—not undermine—confidence in retirement services.
blending AI with human expertise
Artificial intelligence is playing a pivotal role in transforming retirement contact centers. Solutions like Infosys CORTEX leverages AI, NLP and ML to transcribe conversations in real time, extract key insights, and update workflows automatically. These capabilities reduce manual effort, lower error rates, and enhance service consistency. Advanced tools such as sentiment detection and language translation further support agents in managing large volumes efficiently.
Yet, technology alone cannot meet the full spectrum of participant needs. Retirement queries often involve high-stakes financial decisions, personal emotions, and regulatory nuance. Human agents bring empathy and contextual judgment to these conversations—qualities that AI cannot fully replicate. In such scenarios, agents offer reassurance, clarify complex policies, and guide participants through sensitive decisions.
The most effective model combines AI for speed and consistency with human expertise for empathy and interpretation. This hybrid approach ensures routine issues are resolved quickly while complex interactions receive the attention and nuance they require.
key product highlights:
- AI-Driven customer engagement: Makes communication more purposeful, Predicts the intent of real-time customer conversations and Proactively senses caller sentiment in near real-time and provides options for agents to intervene, de-escalate, and mitigate negative situations.
- Agent empowerment and learning: Helps agents make better, faster decisions and provides a no-code configuration driven studio to create realistic contact center audio conversations and scenarios for agent training. Also provides a conversational AI interface for agents to search enterprise knowledge.
- Increased operational efficiency: It is not only limited to streamlining operations through automated processes and effectively improves service to sales conversion but provides conversational analytics to derive values, KPIs, and insights from conversations (calls and chats), including sentiment trends, pain points, opportunities, and improvement recommendations. Most importantly, it provides an automated call summary with issue, cause, and resolution details upon call completion, saving time and effort for agents.
- Cloud first microservice architecture: The foundation is built on a modular microservices-based architecture with cloud computing power (e.g., integrating with Google Cloud Contact Center AI services and AWS services like Amazon Connect and Amazon Bedrock) to achieve scalability and seamless integration.
- Language and ascent neutralization: The platform utilizes powerful language translation models to translate customer queries in non-English languages to English in real-time which helps in overcoming language barriers. Not only that, it also extracts and analyzes microdata from interactions, converting it into insights for real-time action.
looking ahead
As participant expectations shift and service volumes grow, retirement contact centers must evolve. The future lies in tightly integrated operations that blend intelligent automation with secure, participant-focused support. Automating high-frequency queries and enabling seamless transitions between digital and human channels will be essential for operational efficiency and customer satisfaction.
Security and trust will remain paramount. Contact centers must ensure robust data privacy, transparent AI decision-making, and continuous oversight to maintain participant confidence. Real-time insights and contextual support will empower agents to respond with greater accuracy and speed.
Providers that adopt AI strategically, modernize their infrastructure, and design services around participant needs will be better positioned to deliver scalable, resilient, and engaging support models. Future success will rest on the ability to combine digital maturity with operational empathy.
How Infosys BPM can help: At Infosys BPM, we help retirement service providers reimagine their contact center operations. Our solutions combine AI, process expertise and human-centered design to deliver scalable, compliant and engaging participant experiences.
We believe in building a future where technology and people work in harmony—to create trusted, resilient and high-performing service models.
Connect with our team today!
