The $12 Million Question
A Fortune 500 energy company made a major investment in AI by rolling out Microsoft Copilot to thousands of employees. The initiative represented a multi-million-dollar commitment intended to transform how the workforce operated.
Six months in, the CFO walked into a leadership meeting and asked:
“What exactly are we getting for this investment?”
The licenses were active and the rollout was complete. But real usage was flat, and the productivity gains promised in the business case remained largely theoretical.
This was not a technology failure. The software worked as expected. The real problem was visibility. The organization owned a powerful tool but had no clear way to measure whether it was actually delivering value or why adoption was lagging.
The Problem With Native Reporting
Microsoft’s built-in dashboards showed usage numbers, but those numbers raised more questions than they answered.
Were employees actually integrating Copilot into their daily work, or just experimenting with it occasionally? The dashboards could not provide that level of detail. Which teams had successfully adopted it, and which were struggling? What separated someone who had transformed their workflow from someone who had tried it twice and stopped using it?
The data existed, but it was buried within aggregated reports and simplified charts that looked acceptable in executive presentations but could not support real operational decisions.
Leadership faced a difficult choice. They could continue investing in broad training programs and hope adoption improved, or accept that a portion of their AI investment might never deliver the expected return.
The larger issue was the opportunity cost. Across the workforce, meaningful productivity gains were likely possible. However, without visibility into real usage patterns, those gains remained locked inside the organization.
Looking at the Problem Differently
When TechWish became involved, we started with a simple observation that seems obvious in hindsight: departmental averages are almost useless for understanding adoption.
Consider the example of a finance department. It may include analysts working in spreadsheets, controllers reviewing reports, strategists building models, and administrators coordinating meetings. These roles operate very differently and would naturally use Copilot in very different ways, if they used it at all. Aggregating their behavior into a single departmental average reveals very little about how the tool is actually being adopted.
To address this, we built a different approach. Our AI Adoption Analytics Platform does not simply track whether employees use Copilot. Instead, it maps usage against what we call Job Archetypes, behavioral profiles such as the Researcher, the Strategist, the Communicator, or the Coder. These archetypes cut across traditional organizational structures and group employees based on how they actually work.
Once adoption is viewed through this lens, meaningful patterns begin to emerge. Researchers in one part of the organization may be using Copilot heavily, while researchers in another group may not be using it at all. Strategists across multiple teams may struggle because no one has demonstrated the Copilot features most relevant to their workflow.
Under the Hood
Three core data streams power the platform.
First, we connect to Microsoft Graph APIs to capture transaction-level Copilot usage signals. Instead of simply tracking logins, the platform analyzes how Copilot is actually used across applications, what types of interactions occur, and how frequently employees engage with the tool. This creates a much richer picture of usage behavior than traditional reporting dashboards.
Second, we integrate organizational context from systems such as Workday or other HR platforms. This adds meaning to the usage data by linking it to roles, teams, business units, and reporting structures. With this context in place, Copilot activity can be understood in relation to how people actually work inside the organization.
Third, the platform tracks enablement activities such as training sessions, workshops, office hours, and targeted outreach. This closes the loop between adoption insights and enablement actions, making it possible to see which initiatives genuinely improve usage and which ones have little measurable impact.
When these data streams are combined, leadership can answer questions that were previously difficult to evaluate. Which job archetypes benefit most from Copilot? Where are adoption gaps emerging? Which training initiatives actually improve usage patterns?
What We Actually Did
Data only creates value when organizations act on it. Once the platform surfaced usage insights, the next step was translating those insights into targeted action.
Identifying internal champions: The platform highlighted employees who had already integrated Copilot effectively into their daily workflows. These individuals became internal champions who could demonstrate practical use cases and share proven approaches with peers.
Capturing successful workflows: Working with these champions, the team documented real examples of how Copilot supported everyday work. Instead of generic tutorials, these playbooks focused on specific workflows, prompts, and use cases aligned with different job archetypes.
Delivering targeted enablement: Training programs were aligned with job archetypes rather than delivered as broad, generic sessions. Researchers, strategists, analysts, and communicators each received enablement tailored to how they used Copilot in their daily work.
Measuring adoption and refining the approach: Because the platform continuously measured usage patterns, the organization could evaluate which enablement initiatives produced meaningful adoption gains. Successful interventions were expanded, while less effective programs were adjusted or discontinued.
Over time, this created a repeatable model for driving AI adoption across the workforce.
What Happened
Within 90 days, active Copilot utilization increased by 44 percent, reflecting sustained engagement rather than occasional experimentation.
Within six months, the estimated productivity return associated with the analytics platform exceeded 20×, based on interaction analysis and modeled productivity indicators. Leadership now had a defensible way to connect AI investment with measurable outcomes.
Additional outcomes included:
50 million Copilot interaction events analyzed in the first 30 days, providing a comprehensive view of enterprise AI usage
- Up to 95 percent reduction in manual reporting effort through automated executive dashboards
- A centralized, governed view of adoption trends across the organization
Perhaps the most important outcome was strategic. Copilot shifted from being viewed as a cost center to becoming a measurable productivity platform with clear role-based value drivers.
When leadership evaluates future AI initiatives, they now have a structured framework for understanding how adoption translates into business impact.
The Bigger Picture
This scenario is increasingly common across enterprises adopting generative AI. Organizations invest in powerful tools but struggle to translate deployment into measurable productivity gains.
In many cases, the technology itself is not the issue. Tools such as Copilot are capable of delivering meaningful productivity improvements. The challenge lies in understanding how different roles incorporate these tools into their daily workflows.
Vendor dashboards typically focus on surface-level activity metrics. They show that software is being used, but they rarely provide the insight required to understand how usage connects to real business outcomes.
Bridging that gap requires more than reporting. Organizations need visibility into role-based usage patterns, the ability to identify successful workflows, and a structured approach to enabling adoption across the workforce.
For this energy enterprise, that shift transformed a stalled AI investment into a scalable capability. The question many organizations now face is straightforward:
How much unrealized value exists within their own AI deployments?
About TechWish
TechWish is a Microsoft Solutions Partner specializing in AI adoption and workforce intelligence. Our AI Adoption Analytics Platform helps enterprises move beyond surface-level metrics to understand how AI actually creates business value.
We work with Fortune 500 organizations across energy, financial services, healthcare, and manufacturing to translate AI investments into measurable productivity gains.
Interested in unlocking the full value of your AI investments?
Leave a Reply