A Structured Approach to AI Implementation Ensures Success

By: Wayne Leipold, Segal

Public retirement systems are moving AI from experimentation to execution. This article highlights how thoughtful planning, training, and project discipline can develop an AI strategy that helps to produce real results without relying on a single tool or vendor.

Artificial intelligence (AI) has quickly become part of everyday work for many retirement system professionals — summarizing documents, drafting communications or accelerating analysis. But moving from individual experimentation to organization-wide value requires more than access to AI tools. It requires structure.

One public retirement organization recently walked this path, beginning with curiosity about AI’s potential to improve productivity and decision-making — particularly within its technology and project delivery teams — and ending with a coordinated portfolio of strategic initiatives. This experience illustrates how retirement systems can responsibly adopt AI while maintaining the accountability and transparency their members expect.

What followed was not a single “AI project,” but a sequence of coordinated efforts — each building on the last — to ensure AI adoption was useful, well-governed and sustainable.

Start With Strategy for Clarity

The work began with the development of an organizational AI strategy. Rather than focusing on specific technologies, the strategy established three guiding principles:

• AI should augment human judgment.
• AI use must align with organizational values.
• Data privacy, security and transparency are foundational, especially in a public-sector retirement environment.

A key outcome of this phase was the development of a proposed “Responsible Use of AI” policy. The policy translated broad concepts into practical expectations around acceptable use, human oversight, training requirements and accountability. By putting guardrails in place early, the organization created space for innovation while reducing uncertainty for staff and leadership alike.

From a management perspective, this step mattered because it was intended to reduce risk. Staff knew what was allowed; leaders knew how AI would be governed; future projects could move forward without revisiting foundational questions each time.

Build Confidence Through Education

With strategy and policy established, the organization’s attention turned to people. AI training sessions focused on building shared understanding were delivered to staff and leaders. The sessions addressed what generative AI does well, where it falls short and how it can be used responsibly in day-to-day work.

This training helped staff and leaders gain confidence that AI could be used safely and effectively. And the workforce began to identify meaningful opportunities to integrate AI into existing workflows — rather than experimenting in isolation.

Training became the bridge between high-level strategy and practical execution.

Move From Ideas to Action With an AI Tactical Blueprint

Only after strategy, policy, and training were in place did the organization move into implementation. That transition was formalized through an AI tactical blueprint, which is a practical roadmap for turning ideas into action.

The blueprint identified a small set of pilot initiatives, each clearly defined with objectives, success measures, dependencies, and governance. Rather than attempting broad transformation, the organization focused on targeted opportunities where AI could deliver near-term value with manageable risk.

Initial pilots deployed AI agents to support existing work processes. One pilot enabled executive reporting related to the software development lifecycle. A second was created to provide information to team members who were going through an organizational change. The third is an operational knowledge management tool using AI agents for survivor benefit processing teams.

Each initiative followed familiar project-management disciplines: defined scope, phased rollout, stakeholder engagement, and clear ownership.

By introducing it through these specific pilots, AI started to become part of how work gets done.

Lessons That Travel

This organization’s experience provides a broader lesson for public retirement systems: Successful AI adoption is less about finding the “right” tool and more about sequencing the work correctly.

Start with strategy and guardrails. Invest in training to Three Practical Takeaways for Retirement Systems build confidence and trust. Then move into execution 1. Treat AI as a program, not a product. Policies, using disciplined project management and clearly education and training, and governance matter as defined pilots. This approach allows organizations much as the technology. to learn, adapt and scale — without losing sight of 2. Start small and stay focused. Carefully selected accountability or the necessity of appropriate third-party pilots build momentum without overwhelming staff. risk management and vendor due diligence activities. 3. Keep humans in the loop. AI works best when it supports professional judgment. AI may be new, but the fundamentals of managing change are not.

About the author: Wayne Leipold, PMP, is a Senior Consultant in Segal’s Administration and Technology Consulting practice. He has more than 25 years of consulting experience implementing complex software solutions, process improvements and operations management for clients, including public pension systems.