CLEVERPENS

Lingap Ugnay is extending public services to communities that have been having challenges with productivity. The company suffered from what some experts call “shadow competence,” the skills or contributions of employees that go unrecognized or underutilized. Management wanted to provide better service. They believed that with increased speed and capacity, staff would be able to help more people with the same resources. To tackle this, Lingap Ugnay created Dahon, an AI-powered productivity tracking system. To address this, Lingap Ugnay created Dahon, an AI-powered productivity tracking system.

The system generated a productivity score from 0 to 100 for every employee. The score was calculated using data such as the following:

• Number of cases completed
• Response time
• Tasks completed
• Attendance
• Work logged in the organization’s systems
• Customer feedback
• Time spent on particular activities

The operations director was enthusiastic and said, “For the first time, we will have an objective way of knowing who is productive and who needs to improve.” The management team began using Dahon in monthly performance reviews.

The First Surprise

After three months, the dashboard showed that some employees consistently had very high productivity scores. Others had much lower scores. Management initially assumed that the employees with high scores were the organization’s most productive ones. But the operations manager, Clare, became curious. She noticed that some of the employees with the lowest scores were actually spending considerable time outside the office for fieldwork.

They were visiting communities, meeting local stakeholders, resolving problems in the field, conducting outreach activities, and helping people who had difficulty accessing the organization’s services. The systems used by Dahon did not record much of this work.

One employee, David, had a productivity score of only 62. Yet Clare knew that David had spent much of the month working in several communities where service access was particularly difficult. His work had resulted in several long-term improvements. But Dahon saw something different:

• Fewer recorded cases.
• Fewer system transactions.
• More time away from the office.

The AI therefore interpreted David’s activity as relatively low productivity.

The Sick Leave Problem

Then another issue emerged. Jean, one of the organization’s experienced employees, became ill and took an approved sick leave for two weeks.

During that period, she obviously completed no cases. Her productivity score flattened and subsequently fell. When the monthly performance report was generated, Jean appeared to have experienced a significant decline in productivity.

Her manager initially asked, “What happened to Jean’s performance?”

The answer was simple: She was on approved sick leave. The AI had not necessarily made a calculation error. It had correctly observed that Jean had completed fewer tasks. However, the context behind the numbers was not provided. A reduction in recorded output did not necessarily mean a reduction in employee performance.

The Difficult-Work Problem

Clare then compared two employees. Johnny had completed 120 cases and received a productivity score of 95. In contrast, Anna had completed only 65 cases and received a score of 73.

At first glance, Johnny appeared considerably more productive. But Clare examined the cases more closely. Johnny had mostly handled routine requests that took only a few minutes to resolve. Anna had been assigned some of the organization’s most complicated cases. Many required investigation, coordination with other departments, and communication with external stakeholders. Anna had completed fewer cases, as each case required substantially more work.

The AI had measured the number of outputs. It had not fully captured the complexity or value of the work.

The Unintended Consequence

Something even more concerning began to happen. Employees learned how Dahon worked. They discovered that completing more measurable tasks increased their scores. Some employees therefore began choosing easier tasks whenever possible. Others became reluctant to spend time helping colleagues because mentoring did not significantly increase their productivity score.

Some employees avoided complicated cases because these took longer and could reduce their apparent productivity. The organization had unintentionally created a new incentive: “Maximize the score.” But maximizing the score was not necessarily the same as maximizing organizational performance.

The Operations Director’s Dilemma

The operations director called Clare into a meeting.

Director: The system was supposed to improve productivity. Why are we now questioning it?

Clare: Because we may be measuring what is easy to count rather than everything that creates value.

Director: But surely the numbers don’t lie.

Clare: The numbers may be accurate. But an accurate number can still be an incomplete measure of performance.

The room became quiet. Clare continued, “The system tells us how many tasks were recorded, how quickly they were completed, and how employees performed against certain indicators. But it doesn’t necessarily tell us why the numbers look the way they do.”

She gave several examples, like that of David’s score being low because he was doing field work. Jean’s score fell because she was on approved sick leave. Anna’s score is lower because she was handling complex cases. Some employees are changing their behavior because they know what the AI rewards.

The director realized that the problem was bigger than the AI itself. The organization had never clearly answered a fundamental question: What do we actually mean by productivity?

Rethinking Productivity

Lingap Ugnay decided not to abandon Dahon. Instead, the operations team redesigned its approach. They recognized that a single number should not represent productivity. The new performance framework included four dimensions.

1. Quantitative Performance

What can reasonably be measured?

• Output
• Timeliness
• Accuracy
• Work completed
• Resource utilization

2. Quality and Effectiveness

Did the work actually achieve its intended purpose?

• Quality of service
• Customer outcomes
• Resolution of problems
• Rework
• Long-term results

3. Contribution That Is Difficult to Quantify

What valuable work might the system obscure?

• Community outreach
• Field work
• Mentoring
• Team support
• Relationship building
• Problem-solving
• Innovation

4. Context

What circumstances affected the employee’s measured output?

• Approved sick leave
• Training
• Special assignments
• Field deployment
• Complex workloads
• System failures
• Temporary reassignment

The AI score remained part of the system. But it was no longer treated as the employee’s productivity. It became one source of performance information.

A New Conversation About Productivity

At the next management meeting, the operations director presented the new approach. “We made a mistake,” she said. We thought the question was, “Who has the highest productivity score?” She paused.

“The better question is, does the score accurately represent the value and performance of the work being done?” The operations director continued.

The organization then adopted a principle: Measure what matters, not merely what is easy to measure.

Managers were told to combine AI-generated scores with human judgment and operational context. Employees were also provided an opportunity to explain circumstances that the system could not capture.

The Operations Management Lesson

The experience changed how Lingap Ugnay thought about productivity. The organization realized that productivity is not simply the number of tasks completed. It involves the relationship between:

Inputs → Process → Outputs → Quality → Outcomes

AI can be extremely useful in measuring parts of this process. But AI does not automatically understand the full operational context. This creates AI risk. The risk is not necessarily that the AI is malfunctioning. The risk may be that the organization places too much confidence in an incomplete measurement.

Why This Inaccuracy Is an Operations Management Issue

The case is fundamentally about operations management because the organization is trying to answer questions such as the following:

• How should employee productivity be measured?
• How should resources be allocated?
• How can processes be improved?
• How do you measure performance?
• How do we balance efficiency and quality?
• How do we prevent performance measures from creating undesirable behavior?
• How do we ensure that productivity improvements actually improve organizational outcomes?

AI has become part of the operational measurement system. The operations manager must therefore manage not only the process itself but also the measurement system used to manage it.

The Central Problem

The case ultimately raises a deceptively simple question: If something cannot be easily measured by AI, does that mean it is not productive? The answer is NO.

Some activities are straightforward to count but may have limited value. Other activities may be difficult to quantify but create substantial long-term value. Therefore, measured productivity ≠ necessarily actual contribution. And: A high AI score ≠ necessarily high organizational value.

SUMMARY

Lingap Ugnay implemented an AI-based productivity monitoring system, “Dahon.” It unintentionally produced productivity reports with a bias toward quantitative metrics and away from qualitative input. The organization’s measurements of productivity were rethought because of people cheating the system. A new paradigm was developed to incorporate hard-to-measure contributions, quality, and context.

Discussion Questions

• What is the difference between productivity and measured productivity?
• What types of employee contributions did ProScore fail to adequately capture?
• Was ProScore necessarily “wrong,” or was the productivity measurement system incomplete?
• How could the productivity score unintentionally influence employee behavior?
• What are the risks from using an AI-generated productivity score for decisions around promotions, compensation, or discipline?
• How should operations managers balance the following:

a. Efficiency vs. Effectiveness
b. Quantity vs. Quality
c. Standardization vs. Context
d. AI measurement vs. human judgment

• Should an organization ever rely on a single productivity score to evaluate an employee? Why or why not?

TAKEAWAYS

AI can help operations managers measure, analyze, predict, and improve operational performance. However, the introduction of AI creates a new management responsibility: Managers must ensure that what the AI measures is actually what the organization values. Otherwise, the organization may become very efficient at optimizing the wrong thing. The real AI risk is not simply that the machine might obtain the number wrong. It is that the organization might misinterpret the meaning of the number.

DEFINITION OF TERMS

Shadowed Competence: It is a systematic narrowing of human reasoning induced by AI-generated analysis.

DISCLAIMER

This writing is for educational purposes and encourages discussions and debate. It aims to provoke critical thinking in several domains, including philosophical psychology and ethics. It is concerned with perception, judgment, and decision-making. Any mention is for debate purposes, not an accusation of wrongdoing.

The information is inconclusive. Readers are cautioned to distinguish fact from interpretation, to hold themselves accountable to proper procedure and permitted findings, and to be responsible in their research and discussion. The discussion around learning’s true value can be deeply personal, so take what resonates and explore further. As you do so, you may form your own view about learning’s true value.

RELATED READINGS

Split Decision

RELATED LINKS

• Dashboard That Counted People
• Leadership Replacement Exchange Dynamics
• Business Flow of Funds (BFF)
• Eyesight to Insight
• Plan. Manage. Thrive

© 2026 CLEVERPENS