FOMAT & THE AI PRODUCTIVITY TSUNAMI: WHY AI ISN’T MAKING SOME OF US WORK LESS

By Craig Underwood | collaborationevangelist.com

Net/ Summary: Artificial intelligence can absolutely make us more productive. But for executives, consultants, entrepreneurs, nonprofit leaders and other ambitious “power users,” increased productivity does not necessarily mean increased leisure. It can mean more ideas to evaluate, more opportunities to pursue, more reports to read, more decisions to make and more AI “staff” to manage. The leadership challenge is no longer simply learning how to use AI. It is learning how to manage the tsunami of valuable output AI can create — without sacrificing judgment, relationships, focus or our lives outside work.

 

Figure 1. FOMAT — Fear of Missing Agent Time! A humorous beginning to a very real AI-era management problem.

A few months ago, I created a cartoon with AI called FOMAT: Fear of Missing Agent Time! In the cartoon, I awaken in a panic at 2:47 a.m. with a horrifying thought: “OMG, my agent doesn’t have anything to work on!”

OK, the joke is bit absurd. But, for those of us who have learned how to use generative AI well — through iterative prompt engineering, deep research and custom agents — it may also feel disturbingly familiar.

I use ChatGPT, Gemini and, most prominently, Claude Cowork, where I have begun creating my own custom agents. I also use a Plaud device to record, transcribe and summarize meetings. These tools have not made me less engaged in my work. They have increased my energy and engagement and expanded what I believe I can accomplish.

That is exhilarating. It is also a little exhausting.

Because once you discover that an AI agent can research a market, analyze a strategic question, summarize a meeting, develop a donor profile, create a first draft, recommend next steps or identify ten new opportunities before you have finished your coffee, your problem changes.

The question is no longer: “How can I get this work done?” The new question becomes: “How do I possibly keep up with all the valuable work my AI tools can now help me generate?”

Welcome to the AI Productivity Tsunami.

AI IS MAKING MANY OF US MORE PRODUCTIVE. THAT DOESN’T MEAN WE ARE WORKING LESS.

There is increasingly strong evidence that generative AI can improve productivity.

In a controlled experiment involving professional writing tasks, researchers Shakked Noy and Whitney Zhang found that people using ChatGPT completed tasks approximately 40% faster while producing work rated 18% higher in quality.[1]

In another major workplace study, researchers Erik Brynjolfsson, Danielle Li and Lindsey Raymond examined the use of a generative AI assistant by more than 5,000 customer-support agents. Workers using AI resolved approximately 15% more customer issues per hour.[2]

Research conducted with Boston Consulting Group consultants found similar benefits — but also an important warning. Consultants using GPT-4 performed substantially better on tasks within AI’s capabilities, but performed worse when they relied on AI for tasks outside what researchers called its “jagged technological frontier.”[3]

In other words, AI can help us do more, faster and sometimes better. But productivity does not automatically turn into free time. For ambitious people, productivity frequently creates more ambition.

A ten-hour research project that once seemed impractical may now become feasible. A strategic analysis that once required weeks may now be developed in days. A promising idea that previously remained on the “someday” list can suddenly become an active project. And every new project creates more material to read, test, edit, prioritize, validate, and act upon.

Figure 2. Separate studies find meaningful productivity and quality gains from generative AI; they do not measure a single combined effect.

MY OWN EXAMPLE: AN AI AGENT THAT CREATES OPPORTUNITY — AND MORE WORK

I currently help lead fundraising work for Sportsmen’s Tennis & Enrichment Center (STEC), a Boston nonprofit that provides tennis, academic support, health and wellness programs and community opportunities for thousands of young people and their families.

One important assignment is identifying companies and foundations in Greater Boston that may be strong prospective donors or sponsors.

Historically, researching hundreds of companies from the Boston Business Journal’s Book of Lists would require enormous staff time or outsourced to a charitable consulting firm: identifying companies, researching corporate citizenship priorities, determining whether they give directly or through foundations, reviewing charitable contributions and tax filings when available, assessing connections to youth development or Boston communities, and prioritizing outreach.

Today, I have created an AI-enabled workflow to help research and evaluate more than 500 companies and prioritize those most likely to contribute to STEC.

That is a remarkable capability. But it does not mean I now have nothing to do. Quite the opposite.

The agent can help identify prospects, summarize evidence, assemble information and suggest rankings. But a human still must decide whether the information is accurate and current; which opportunities deserve immediate attention; what kind of contribution to request; who might know someone inside the company; and how to convert intelligence into genuine human connection and financial support.

The agent produces information. The leader must still exercise judgment. And when the information is excellent, the temptation is to pursue more of it. That is exactly how the tsunami begins.

Figure 3. The AI Agents’ Productivity Tsunami: when valuable output creates a new prioritization challenge.

RESEARCH NOW SUPPORTS THE TSUNAMI HYPOTHESIS

My cartoon was intended to be humorous. Then I discovered research suggesting it may also be descriptive.

In February 2026, Harvard Business Review published an article by Aruna Ranganathan and Xingqi Maggie Ye of the University of California, Berkeley Haas School of Business. Based on an eight-month ethnographic study at a roughly 200-person U.S. technology company, their in-progress research found that generative AI did not simply free up time. Employees worked faster, broadened the scope of tasks they undertook and extended work into more hours of the day, often without being instructed to do so.[4]

The explanation is both simple and powerful: AI made doing more feel possible, accessible and rewarding. That is precisely the dynamic many AI power users are beginning to experience.

Figure 4. Recent Berkeley Haas research featured in Harvard Business Review closely mirrors the “productivity tsunami” hypothesis.

Microsoft’s 2025 Work Trend Index provides additional context. Its telemetry analysis found that the top 20% of Microsoft 365 users by ping volume received 275 meetings, emails or chats per workday — equivalent to one interruption approximately every two minutes during an eight-hour workday.[5]

Now add AI agents that can continuously create additional analyses, drafts, ideas, recommendations and action items. AI may help close the capacity gap. But unless leaders deliberately redesign how they work, it can also pour more water into an already overflowing ocean.

Figure 5. Microsoft finds severe input overload among its most highly pinged Microsoft 365 users.

To be clear, this is not everyone’s experience. OECD research on small and medium-sized enterprises using generative AI found that 32.7% reported decreased workload, while 11.8% reported increased workload.[6]

My hypothesis is narrower: for highly motivated leaders and professionals who become skilled AI users — especially those creating their own agents — AI may initially increase rather than decrease total work because it expands the frontier of worthwhile opportunity.

Figure 6. An important counterpoint: among SMEs using generative AI, more reported decreased workload than increased workload.

WHAT ABOUT JOB LOSS?

There is another side of this discussion that leaders should not ignore.

Some highly repetitive, process-oriented and entry-level roles are likely to shrink or change substantially as AI becomes better at routine research, summarization, document preparation, basic customer support, scheduling, data entry and administrative work.

The International Labour Organization has found that clerical occupations remain among those most exposed to generative AI transformation.[7] The World Economic Forum’s Future of Jobs Report 2025 similarly identified clerical and administrative positions — including data entry clerks, bank tellers, cashiers and administrative assistants — among the roles employers expect to decline most quickly through 2030.[8]

I believe leaders have the responsibility – some would say the moral imperative – to help employees learn how to use AI responsibly and efficiently and to move toward higher-value work involving judgment, relationships, creativity, influence, and accountability.  This is especially important for nonprofit leaders of work-force development and college access/ success organizations, and those whose missions include preparing those they serve for jobs of the future.  Imagine the different impression a young entry-level job or internship applicant would leave on a prospective employer when answering the question “Do you use AI?” between:

“[My training organization] has provided responsible AI literacy training and prompt engineering to all of its students.”

And

“I don’t use AI much at all.” Or even worse, “We are prohibited from using AI.”

 

THE SOLUTION: SURF. PRIORITIZE. IMPACT.

In the upper-left corner of my Productivity Tsunami cartoon is a simple phrase: Surf. Prioritize. Impact. I think this may be the practical leadership framework we need.

1. SURF: USE AI ENTHUSIASTICALLY — BUT PURPOSEFULLY.

The answer is not to run away from AI. It is too valuable. These tools can help leaders think better, research faster, communicate more clearly and identify opportunities that might otherwise go undiscovered.

But surfing a wave is different from letting it drown you.

Leaders should be explicit about what they want AI to help accomplish. Or, to borrow a phrase from Steven Covey’s The 7 Habits of Highly Effective People, “begin with the end in mind.”  Create agents around important outcomes, not simply because creating another agent is interesting. Keep a running “AI opportunity parking lot” for intriguing ideas that do not deserve immediate action.

2. PRIORITIZE: SCHEDULE MEETING TIME WITH YOUR AI OUTPUT.

If you manage human colleagues, you do not expect them to slide dozens of excellent reports under your door every night and assume you will process them all by morning. The same discipline should apply to AI agents.

Schedule regular “meeting time” to review AI-generated work: a weekly (or daily) review of agent findings and recommendations; a clear decision rubric for what moves forward, what gets delegated, deferred or discarded; limits on active AI-enabled projects; and a requirement that all AI research and other generated content receive human verification and reflection.

The ability to generate more options makes disciplined prioritization more important, not less.

3. IMPACT: PROTECT THE WORK ONLY HUMANS CAN DO.

AI can produce a donor prospect analysis. It cannot build trust with a prospective sponsor over coffee. AI can summarize a leadership challenge. It cannot look a struggling colleague in the eye and ask the compassionate question that changes a relationship. AI can generate executive coaching frameworks and meeting notes. It cannot replace the human wisdom, accountability and courage required to make a difficult choice.

Leaders should protect time for relationships, reflection and judgment — the very human activities that make AI-generated intelligence meaningful. And we should place boundaries around our AI-enabled work lives.  We must make sure that those of us who become highly productive through AI do not confuse increased output with increased impact — or silently turn every available evening, weekend, and quiet moment into additional time supervising our digital workforce.  Given my passion for and excitement about “living on an AI vertical learning curve, “ my wife and I often have “AI discussion timeouts,” where I am not allowed to talk about it – much less, fire up Claude on my laptop to show her the latest cool thing we generated together.

Not every new research question needs to be launched tonight. Not every good idea requires immediate development. Not every idle agent needs a new assignment before bedtime.

Sometimes the most productive instruction we can give our AI agents is: “Good work. We will review this tomorrow.”

THE HUMAN STILL HAS TO CHOOSE THE SHORE

Artificial intelligence is not simply reducing the burden of work. For many of us, it is revealing a much larger universe of work that could be done.

That is not necessarily bad news. It means that a small consulting firm, nonprofit leader, entrepreneur or executive can now access analytical, creative and research capacity once available only to organizations with large staffs and large budgets.

But abundance creates a new leadership obligation. The best leaders in the AI era will not be the people whose agents generate the most reports, the most ideas or the most activity. They will be the people who know which waves to ride, which opportunities to ignore, which decisions require human judgment and when to step out of the ocean entirely.

Because the goal of AI is not to keep every agent busy. The goal is to help human beings create meaningful impact — while still leaving enough time and energy to be fully human.

Footnotes

[1] Shakked Noy and Whitney Zhang, “Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence,” Science 381, no. 6654 (2023): 187–192. Finding cited in the article abstract: average time decreased 40% and quality rose 18%.

[2] Erik Brynjolfsson, Danielle Li and Lindsey Raymond, “Generative AI at Work,” The Quarterly Journal of Economics 140, no. 2 (2025): 889–942. Published article reports a 15% productivity increase measured by issues resolved per hour.

[3] Fabrizio Dell’Acqua et al., “Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality,” working paper, 2023; subsequently published in Organization Science.

[4] Aruna Ranganathan and Xingqi Maggie Ye, “AI Doesn’t Reduce Work—It Intensifies It,” Harvard Business Review, February 9, 2026; Laura Counts, “AI Promised to Free Up Workers’ Time. UC Berkeley Haas Researchers Found the Opposite,” Haas News, February 18, 2026.

[5] Microsoft WorkLab, “Breaking Down the Infinite Workday,” June 17, 2025; Microsoft, 2025 Work Trend Index Annual Report. The 275-per-day figure is based on the top 20% of Microsoft 365 users by ping volume received; the two-minute figure is calculated over an eight-hour workday.

[6] Organisation for Economic Co-operation and Development, Generative AI and the SME Workforce: New Survey Evidence (2025), Figure 3.5.

[7] International Labour Organization, Generative AI and Jobs: A Refined Global Index of Occupational Exposure (2025).

[8] World Economic Forum, Future of Jobs Report 2025 (2025).

Research Bibliography

Brynjolfsson, Erik, Danielle Li and Lindsey Raymond. “Generative AI at Work.” The Quarterly Journal of Economics 140, no. 2 (2025): 889–942.

Dell’Acqua, Fabrizio, et al. “Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality.” Working paper, 2023; subsequently published in Organization Science.

International Labour Organization. Generative AI and Jobs: A Refined Global Index of Occupational Exposure. 2025.

Microsoft. 2025 Work Trend Index Annual Report. 2025.

Microsoft WorkLab. “Breaking Down the Infinite Workday.” June 17, 2025.

Noy, Shakked and Whitney Zhang. “Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence.” Science 381, no. 6654 (2023): 187–192.

Organisation for Economic Co-operation and Development. Generative AI and the SME Workforce: New Survey Evidence. 2025.

Ranganathan, Aruna and Xingqi Maggie Ye. “AI Doesn’t Reduce Work—It Intensifies It.” Harvard Business Review. February 9, 2026.

University of California, Berkeley Haas School of Business. “AI Promised to Free Up Workers’ Time. UC Berkeley Haas Researchers Found the Opposite.” February 18, 2026.

World Economic Forum. Future of Jobs Report 2025. 2025.