By Jerrod Weiss

On March 5, 2026, more than 100 professionals from across the retirement industry gathered in New York City for a forum on one of the most talked-about topics today: artificial intelligence. The audience was largely made up of engineers, product leaders and compliance professionals from recordkeeping firms. Early on, it became clear this wasn’t a typical networking event – it was an active exchange of ideas, perspectives, and, at times, uncertainty.

There is clear excitement across the industry but also hesitation. Engineers are eager to push AI capabilities forward as quickly as possible. Product teams are focused on finding the right timing and strategy. Meanwhile compliance teams are working to understand how to govern and regulate these tools.

There is no shortage of perspectives on AI’s impact on the 401(k) industry. With adoption accelerating so quickly, much of what was discussed may feel outdated within a year. Still, a few key insights stood out and centered on three themes.

1. How AI is showing up today
2. What plan sponsors and retirement consultants are asking
3. Governance and risk management

How AI is Showing Up Today

• A Fintech RIA firm has established an internal AI Ethics Committee to guide responsible use.
• One asset manager noted the challenge of building AI solutions for all participants—high-net-worth individuals prefer a human touch, while smaller-balance participants often lack trust in the technology.
• A relationship manager shared a story from a competitive finalist presentation, where a prospect used AI to evaluate providers—including their data security. The results were so inaccurate that the prospect disclosed the bad information to all competing firms.
• When attendees were asked how many had used AI for financial advice, roughly two-thirds raised their hands—suggesting even higher use among the general population.
• Providers are using AI to validate that plan document provisions align with recordkeeping systems, enhancing quality control.
• AI-driven participant advice tools are being used for contribution rates and investment strategies, with built-in participant validation before execution.
• Firms are leveraging inbound participant data to power more personalized outbound communications through AI.
• One provider is piloting AI-assisted call center technology, where representatives respond through an AI interface that refines answers in real time and can even adjust voice tone and dialect.

What Sponsors and Consultants are Asking

• Certain industries, like education, are especially sensitive to AI. This issue is likely to expand across sectors with employees threatened by the integration of AI in the workplace.
• RFPs are beginning to include requests for the ability to opt out of AI-driven processes at both the plan and participant level.
• One client questioned whether workforce reductions driven by AI would lead to lower fees. The reality: the technology remains expensive to implement and maintain.
• Sponsors are asking for real-time insights on how to maximize the value of newly implemented AI capabilities.
• There is growing focus on defining appropriate compliance guardrails.

Governance and Risk Management

A key question emerged: If AI makes a mistake, who is responsible?
Most agreed AI should be treated like an employee:

• What went wrong?
  • Internal programming or engineering?
  • External or poor-quality data?
  • The Alaska Airlines chatbot case was cited as an example, where the company was held accountable for inaccurate information provided by its AI chatbot.

Managing bad data remains a longstanding challenge that can be amplified significantly by AI. Traditional compliance review processes struggle to keep up with AI-generated material, particularly content like videos. AI is evolving faster than traditional control frameworks (such as annual SOC audits) can adapt. Weak data foundations only widen that gap. Insurance providers are continuously reviewing underwriting standards as AI adoption increases.

After recapping the Forum with the American Trust Retirement leadership team, one comment stood out: “AI amplifies the quality of the environment it is placed in, good and bad.” Successful AI implementation still depends on one foundational element: quality d