Worker Impact Is the Common Denominator of Every AI Transformation – And the Best Early Indicator of Success

---

AI investments are expected to top $2.5 trillion by 2033, but for all its promise, organizations are still struggling to measure its impact. While 79 percent of leaders know they need AI to stay competitive, 59 percent worry about quantifying productivity gains.

This is the economic reality of AI. With investments that can run into the tens of millions of dollars, AI tools (and the leaders who bring them to the table) are under enormous pressure to demonstrate their cost savings or productivity gains as quickly as possible. Unfortunately, that’s not the general timeline for most AI solutions, which often take years to prove their worth.

The solution lies in work-level data. In this piece, we’ll explain how to classify your AI investment based on the work it impacts and how to measure its impact so you can get a clear assessment within months – not years – of whether the investment is paying off.

Classify Your AI Investment: Highly Defined vs. Open-Ended Work       

To measure an AI tool’s impact on work, you have to first define the nature of the work it’s meant to impact. In our experience, the biggest differentiator is whether a worker’s tasks are clearly defined or open-ended.

AI projects for workers with well-defined roles tend to be those that carry the highest expectations for productivity improvements. Examples include…

• An AI-enabled chatbot to assist call center agents.
• A coding tool to speed up the work of software developers.
• A paperwork-reducing tool aimed at reducing attrition among healthcare workers.

These transformations come with very clear goals and substantial investments, putting leaders under intense pressure to know whether their AI tools are working as quickly as possible.

Meanwhile, the second category of AI transformation involves tools designed to increase the productivity of general knowledge workers, such as…

• ChatGPT to help with crafting memos and email messages. 
• CoPilot to assist with a variety of administrative tasks.

These transformations tend to be a lighter lift with less pressure, a smaller investment, and a lower demand for strict bottom-line results.

Measure AI’s Impact on Employee Work

Regardless of the nature of the work AI is disrupting, the way to evaluate the investment is to measure the work. But how do you do that? The answer starts with work friction.

Work friction is anything that gets in the way of a worker doing their job, including people, processes, and technology. A broken headset, for example, or a professional development approval process with too many layers.

To measure work friction, you have to survey workers directly about the moment-to-moment reality of their work. Unlike employee experience surveys, which ask how workers feel about various aspects of their jobs, work friction surveys aim to identify what’s actually happening as they go about their days.

When measuring the impact of AI tools on work friction, you can ask survey questions about how the tool impacts work moments: Is it making them better? Worse? Adding new moments of friction? The answers are your first indicator of whether the AI is delivering a positive ROI.

They do that by highlighting both “success stories” (i.e., where the AI is making work more efficient or workers productive) and problem areas, which gives you a clear indication of where to stay the course and where to adjust.

Best of all, the valuable user feedback you get can help with AI transformations targeting both well-defined and lesser-defined roles. Every AI implementation comes with costs, risks, and questions about whether to expand, alter, or abandon the use of a tool. Work friction data can tell you very clearly if AI is making employees’ work easier or more streamlined.   

A Third Category: AI Tools For Enterprise Services

A third and quickly growing type of AI transformation involves tools designed to assist in providing enterprise services to employees, such as help with payroll or PTO requests. These types of projects generally carry significant cost-savings expectations – as well as high levels of scrutiny from workers. 

Here again, work friction data can provide critical feedback as to whether an AI tool is meeting employees’ needs. For example, we recently worked with a firm that was looking to reduce operational costs and improve the employee experience by investing in a variety of AI chatbots, employee self-service portals, and advanced service management tools.

However, the project was plagued early on by low adoption rates and growing employee frustration, for a number of reasons:

• Processes that seemed straightforward on paper became complex when applied in real-life scenarios.
• Gaps in resources and misaligned systems left employees to solve many problems on their own.
• The new system proved to be less efficient in real life than on paper.

But by studying work friction data to understand exactly where employees were experiencing difficulties with the new system, the firm was able to revise its implementation to take a more targeted approach. This led to streamlined processes and simplified task handovers that resulted in higher employee adoption rates and $2.3 million in annual operational cost savings.

In AI Transformations, Only Employees Can Tell You What’s Working 

The wave of AI will be coming for almost every role within an organization eventually. Even the bluest of blue-collar jobs will be touched by AI at some point in the not-too-distant future. That’s why it’s so important for leaders to figure out how to evaluate these projects on a shorter timeline.

The key is gathering data on work tasks directly from employees. Only your employees can tell you if AI is working. That’s why work friction data is so important. It can serve as a leading indicator for AI success – and provide key guidance for adjusting a deployment that may not be working as expected. Ready to find out more? Get in touch.