Artificial intelligence adoption is often discussed in abstract terms. Pilot programs. Vendor demos. Strategic roadmaps that never quite leave the page. But a recent report from JPMorgan Chase, highlighted by VentureBeat, offers a more grounded example of what large-scale AI use can look like when governance, infrastructure, and organizational trust are aligned. According to the report, roughly half of JPMorgan’s workforce now uses internal AI tools. Not because they were mandated to do so, but because the tools were built to connect directly to real workflows, data, and decision-making needs. For public employee retirement systems, the lesson is less about copying a bank’s technology stack and more about understanding the governance conditions that make adoption possible and sustainable.

Governance Before Gadgets

One of the clearest takeaways from JPMorgan’s experience is that AI adoption followed governance, not the other way around. The firm emphasized clear boundaries around data access, use cases, and internal controls before encouraging widespread use.

For pension systems, this mirrors familiar responsibilities. Trustees are already responsible for overseeing risk, ensuring data integrity, and safeguarding sensitive member information. Any use of AI, whether for reporting, member communications, compliance tracking, or internal analysis, must adhere to the same fiduciary guardrails.

The question for boards is not “Should we use AI?” but rather:

• What data would these tools access?
• Who oversees their use?
• How are outputs reviewed, validated, and explained?

Connectivity as a Strategic Choice

Another core insight from the article is that JPMorgan focused less on flashy AI capabilities and more on connectivity. Tools gained traction because they plugged into existing systems and reduced friction in everyday work.

That principle translates directly to public pension operations. AI tools that sit outside core systems, require duplicate data entry, or lack integration with actuarial, finance, or member platforms are unlikely to gain trust or sustained use.

From a governance perspective, connectivity is not just an IT issue. It is a strategic decision about how information flows through the organization and how decision-makers access consistent, reliable data.

Adoption as a Signal, Not a Goal

Notably, JPMorgan did not set an adoption quota. Usage grew organically as employees found value. That distinction matters for public sector organizations, where forced adoption can raise concerns about transparency, accountability, and unintended risk.

For trustees and administrators, adoption rates should be viewed as a signal, not a success metric. High usage may indicate that tools are genuinely improving clarity and efficiency. Low usage may reveal gaps in training, trust, or governance readiness.

Either way, boards benefit from asking why.

Implications for Public Pension Oversight

Public retirement systems are not investment banks, but they face similar pressures: growing data volumes, increasing reporting requirements, cybersecurity risks, and the need to accomplish more with limited staff resources.

The governance lesson from JPMorgan’s experience is that effective AI use depends less on technological ambition and more on:

• Clear oversight structures
• Defined use cases aligned with the mission
• Integration with trusted systems
• Ongoing human review and accountability

As pension boards continue to modernize their operations, AI will likely become an integral part of that conversation. The strategic challenge is to ensure that those tools support fiduciary discipline rather than distract from it.

In that sense, JPMorgan’s story is less about artificial intelligence and more about institutional readiness. And that is a topic squarely within the oversight responsibilities of today’s pension trustees and administrators.

Questions for Trustees and Administrators to Consider

As Texas public retirement systems consider new analytical, automation, or AI-enabled tools, trustees and administrators may wish to evaluate them through the same governance lens applied to all fiduciary decisions, consistent with guidance from the Texas Pension Review Board.

• Is the purpose clearly defined and documented? Does the proposed tool support an established operational, compliance, or reporting objective consistent with the system’s mission and fiduciary responsibilities?

• Does the use align with existing policies and controls? How does the tool fit within current policies related to internal controls, data security, procurement, and risk management?

• What data inputs are involved, and how are they safeguarded? Does the tool access confidential member, financial, or personnel data, and are safeguards consistent with board-approved data governance and cybersecurity practices?

• Who retains responsibility for decisions? Are trustees and staff maintaining appropriate human oversight, with AI-assisted outputs serving as informational support rather than decision-makers?

• How are accuracy and reliability evaluated? Are there documented procedures for reviewing, validating, and challenging outputs before they are used in reports, recommendations, or board actions?

• Is integration reducing risk or adding complexity? Does the tool integrate with existing systems in a way that strengthens internal controls, or does it create parallel processes that increase operational risk?

• How will performance and risk be monitored over time? What metrics, reviews, or periodic assessments will the board use to determine whether continued use remains prudent and appropriate? 

About the Author: Allen Jones serves as TEXPERS' Director of Communications and Event Marketing. He brings more than two decades of experience in journalism and publication management and now guides the Association's strategic communications. [email protected]