For the past few years, getting employees to use AI was the goal. Companies bought licenses, launched training programs, and watched adoption climb. That was the first hurdle. Now there’s a harder one: Are people actually getting better at using AI? And how do companies know?
One Model AI Impact is designed to answer that question. It brings governed AI activity together with workforce context to help organizations understand what work AI is supporting, how effectively people are using it, and where that use is creating value.
An employee who logs into an AI tool every day might be highly capable. Or they might be using it for the same basic tasks they tried six months ago. An engineering team might have incorporated AI into repeatable workflows that save hours of work, while another team with an identical adoption rate is still experimenting with one-off prompts.
On an adoption dashboard, they can look exactly the same. For technology leaders responsible for increasingly large AI investments, that’s a problem.
Adoption tells you whether employees showed up. It doesn’t tell you whether AI is becoming part of meaningful work, whether employees are building stronger capabilities, or where that use is creating value.
Enterprise AI investment is quickly maturing, but workforce adoption still has a long way to go. Organizations are buying tools, expanding access, and seeing more employees experiment with AI. The harder challenge is turning that early use into consistent habits, stronger capabilities, and more meaningful ways of working.
In One Model’s research of 50 senior technology buyers, 84% of organizations had already moved beyond the pilot stage. Fifty-six percent spend more than $1 million on AI annually, and 94% expect their investment to grow over the next year.
Yet many organizations still struggle to understand what that adoption is producing. Nearly half use employee adoption or usage as a proxy for value, while only 20% have a formal AI ROI framework applied across tools.
Usage matters, but it only answers the first question: Are people using AI? It doesn’t tell leaders whether that use is more valuable. It’s much less useful when the CEO or board wants to know whether those investments are changing how the company performs. And that pressure is already here. Ninety-eight percent of the technology leaders we surveyed have either faced pressure to justify AI spending or expect to soon.
If usage keeps rising, the next question is inevitable: What are we getting better at because of it?
There’s no single behavior that makes someone “good at AI.” Capability is multidimensional. It includes how effectively someone communicates with AI, how efficiently they move from a question to a useful result, how much meaningful work they can delegate, whether AI produces usable outputs, how consistently they use it, and how broadly they apply it across their work.
That means two employees with the same number of prompts can have very different levels of capability. One might use AI occasionally for simple requests. Another might delegate more complex work, incorporate multiple tools, build repeatable workflows, and consistently turn AI interactions into useful outputs.
Those behaviors also look different depending on the work. A software engineer might use AI to debug code, explore alternative approaches, document a solution, or accelerate development. A salesperson might use it to research an account, prepare for a meeting, analyze objections, and tailor follow-up. A finance team might use it to interrogate data, investigate variances, or model scenarios. Simply comparing how many prompts those employees submit tells you very little about how effectively they work with AI.
AI Impact goes deeper than identifying what people use AI for. With prompt-level intelligence enabled, it helps organizations understand patterns in how effectively people work with AI, how much meaningful work they delegate, whether use is becoming habitual, and where stronger practices are emerging. It also provides context on the work AI supports, from research and analysis to coding, content creation, and other tasks.
Instead of asking, “Who uses AI the most?” leaders can begin asking better questions: Which teams are building stronger AI capabilities? What are they doing differently? Where are employees getting stuck? And how can we help more people adopt the practices that work?
Knowing where AI capability is developing, and where people are getting stuck, is useful for leaders. Measurement becomes much more valuable when employees can use it themselves. AI Impact gives individual users a view into their own AI behaviors, including what they’re doing well, where their habits may be costing them time, and how they could work with AI more effectively. Rather than simply assigning someone an adoption or capability score, it provides specific coaching based on how they actually use AI.
For example, an employee might learn that they’re doing a good job giving AI clear constraints and specific instructions, but losing time through a series of fragmented follow-up prompts. AI Impact can surface that pattern, recommend a more efficient approach, and estimate both the time AI is already saving them and the additional time they could recover by improving how they work with it.
That creates a feedback loop traditional AI training can’t provide. A workshop can teach employees general AI techniques and best practices. AI Impact can help them understand their own habits, based on their own AI usage, and give them concrete ways to improve.
As those behaviors change, employees can see whether they’re becoming more effective over time. For technology leaders, that creates an opportunity to move beyond simply measuring AI proficiency. They can give employees a tool to actively improve it.
Individual coaching is one part of the equation. Managers play an equally important role in helping teams turn stronger AI practices into better ways of working.
AI Impact can give managers an aggregated view of where their teams are building capability, where employees may need more support, and which patterns are producing better results. A manager might see that one team is successfully delegating more complex work to AI while another is using the same tools primarily for simple requests. They might also identify a repeatable workflow that is saving one group significant time and could be adapted elsewhere.
Helping one employee get better at AI is valuable. The bigger opportunity is understanding what happens when those improvements spread across the workforce. Most organizations already have employees and teams figuring out what works. One team discovers a workflow that dramatically speeds up a recurring task. A group of power users learns how to get consistently better outputs from a particular tool. Another team finds that a different AI tool is better suited to the work they’re trying to accomplish. Those lessons often stay local.
By connecting governed AI activity with workforce context, AI Impact can help leaders understand how capability differs across people, roles, teams, functions, and types of work. Organizations can also analyze those patterns alongside workforce data such as performance, hiring, movement, cost centers, and other people data to better understand where stronger AI practices are taking hold. That can reveal where stronger AI practices are emerging, where employees may need more support, and where successful approaches could be shared more broadly.
Instead of rolling out generic training and hoping it changes behavior, organizations can make enablement much more targeted. Measure what people are doing. Help them improve. Identify what’s working. Scale those practices across the business. The goal isn’t simply more AI use. It’s better AI use.
Of course, capability isn’t the final destination either. An employee can become exceptionally skilled at using AI without producing a meaningful return for the organization. Ultimately, leaders need to understand what the organization is getting back from its AI investment.
AI Impact helps organizations begin quantifying that return by modeling the value of the time and work enabled by AI and comparing it with the cost of AI activity. That gives leaders a clearer view of where AI is creating value, where friction or wasted spend remains, and how return differs across teams and types of work.
Our research suggests buyers are already thinking this way. Productivity impact was the highest-rated AI measurement challenge among respondents, followed closely by measuring AI ROI. When we asked which capability mattered most in a solution like One Model AI Impact, showing leadership the impact on team and business outcomes ranked first. Tracking productivity impact over time ranked second.
That’s a significant shift from traditional software adoption. The objective isn't to maximize usage. It’s to understand where AI changes the economics or performance of work.
An engineering organization can investigate whether stronger AI capability corresponds with faster delivery. Maybe a customer support team can examine whether AI-supported workflows correspond with faster or more efficient case resolution. Maybe certain workflows become dramatically faster while others show little change at all.
Those are the connections leaders ultimately need to investigate. AI Impact brings AI activity together with workforce context to help organizations quantify value today and build toward an even broader view of impact. By connecting productivity and operational data from systems where work happens, organizations can investigate how AI use relates to delivery, sales, support, hiring, and other business outcomes. From who has AI, to who uses it, to who uses it effectively, to what difference that use makes.
Getting employees to try AI was an important first step. But rising adoption alone can’t tell technology, finance, people, or AI transformation leaders whether their workforce is becoming more capable or whether their AI strategy is creating value.
As investment grows, the standard has to rise with it. Where is your workforce building stronger AI capabilities? Which investments are producing stronger behaviors and measurable value? And, most importantly, where does better AI use translate into better performance?
Those are harder questions than counting active users. They’re also much closer to the questions your CEO, CFO, and board actually want answered.