The Challenge

AECI’s leadership team saw clear potential in generative AI to streamline knowledge-heavy work. But a few key challenges stood in the way.

First, security stakes were high. As a utility provider, any platform adopted would need to meet strict compliance requirements—across legal, technical, and operational dimensions. Second, adoption was uneven. While a handful of early adopters had begun experimenting with AI tools, most employees had only dabbled with ChatGPT, with little consistency or sustained use. And finally, there was no obvious internal owner. Who should take charge of platform selection, write company policy, or lead training for over 700 employees?

The Mandate

Our role was to bring order to that complexity. We were tasked with helping AECI select the right enterprise-grade platform by evaluating OpenAI, Microsoft, Google, and Anthropic against their unique security, cost, and user experience requirements. From there, we needed to design a scalable adoption model that could track progress across divisions, teams, and job levels.

The engagement also called for the launch of an internal Champion program—enabling power users across the organization to lead change from within. And finally, all progress had to be tied to measurable outcomes: time saved, quality gains, and real platform usage before any custom development would be considered.

The Work

We began with platform selection, using a Gartner-style scoring rubric to evaluate over 30 criteria—from SOC 2 compliance and legal indemnities to roadmap alignment and user experience. Through a series of structured deep dives and vendor demos, AECI’s leadership team ultimately selected ChatGPT Enterprise, keeping Microsoft Copilot in view for future integration.

Next, we launched the Champion Program. After a company-wide interest survey, we scored applicants using a rubric and selected a final cohort of twelve employees from a cross-section of departments. These Champions completed a five-part certification program covering advanced prompting, workflow design, AI ethics, security, and change leadership. They now serve as internal coaches—running AI “help desk” in internal forums and modeling practical use for their peers.

Meanwhile, we rolled out training to the broader organization through our three-part AI Accelerator program. We organized employee cohorts of 15 who participated in live, hands-on sessions designed around their real work and data. Training was scheduled strategically to avoid business disruption. A live KPI dashboard, developed in partnership with AECI’s internal team, now tracks adoption across sites and seniority levels using ChatGPT Enterprise analytics.

We continue to lead quarterly workshops focused on strategy, platform shifts, and governance best practices. Adoption scorecards now cascade through the organization, with AI usage and time-saved metrics tied directly to existing OKRs.

Early Momentum

The first six months have shown clear signs of success. All twelve Champions completed the certification program with their own custom GPTs and structured prompt libraries tailored to their daily tasks. Session feedback has been overwhelmingly positive, with 95% of all responses rated “Excellent,” “Very Good,” or “Good” and an average score of 4.2 out of 5 from over 180 ratings.

Perhaps most importantly, the work is already saving time. Champions reported gaining back two to twelve hours each week by applying what they learned. Employees have also responded with enthusiasm. One training attendee from Power Production shared, “That just saved me three days. Incredible.” Another noted, “I was skeptical that AI would help with my everyday work—after this first session, it already has.”

Why It Worked

Three principles guided our approach: focus on adoption before automation, empower internal Champions rather than over-relying on external consultants, and track every step with live, transparent metrics. This model ensured that AI use became a daily habit before any technical build-out, and gave AECI leadership a clear view into usage, value, and impact across the organization.

What Other Utilities Can Learn

Security doesn’t have to slow innovation. With the right evaluation process, even highly regulated industries can move from vendor selection to production-ready tools in under a quarter. Champion programs unlock scale, allowing a small group of trained employees to support hundreds of peers. And when you track the right metrics—usage frequency, time saved, and adoption rate—you can tie your AI investment directly to meaningful ROI.