🔍 Read the full analysis: How The Harvard Study Separates AI Hype From App-Switching Reality on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A Harvard Business School preprint analyzed 103 million application events from 1,017 office workers at eight Indian companies. It found fewer app switches in the ten uses after an AI interaction, but about 200 more switches on days employees used AI; the observational study cannot establish why.
A Harvard Business School preprint examining 103 million application events found a split pattern: workers switched among fewer apps in the ten app uses after an AI interaction, but made about 200 more switches on days they used AI. The study records associations, not a causal effect, so it does not show that AI either improves focus or causes distraction.
Researchers Sumer S. Vaid and Ashley V. Whillans analyzed computer activity from 1,017 employees at eight organizations in India, recorded between November 2025 and January 2026. The data included 56.9 million app switches across 62,693 person-days. Workers switched applications an average of 908.3 times per workday; the median was 814. Those figures count a change to a different application, not a switch between documents within the same app.
The paper’s AI finding depends on the time scale. Looking at the ten app uses after each recorded AI use, the researchers found people used fewer apps, switched less, stayed roughly half a second longer in each app and followed a more predictable sequence. Looking across entire days, however, employees made about 200 more switches on days with AI use, and days with more AI minutes were more fragmented. The authors say the direction of the relationship is unknown: AI could affect how the day unfolds, or workers could use AI when their work is already changing.
The data came from Flowace, a commercial workforce-analytics service whose software records changes in the active computer application, along with timing and other window information. Employers determined which applications counted as productive. The researchers say the data were anonymized before they received them and that a Harvard human-subjects committee approved the protocol. The paper states that it received no external funding and that Flowace had no role in the design or interpretation.
App switching and AI · one moment
What happens around one AI query?
The study zoomed in on every AI use: the ten app uses before it, and the ten after. Up close, work calms down a little. Zoom out to the whole day, and the picture changes.
- 1 · A workday of switchingWorkers changed apps 908 times a day on average. One app held their focus for 27.8 seconds on average, 3.3 seconds at the median.
- 2 · AI comes in at a busy momentPeople used AI a median of 13 times a day, for a median of 7 seconds. Fragmentation peaked in the hour of AI use.
- 3 · The next ten app usesFewer different apps, fewer switches, about half a second longer per app, more predictable. Real, but small.
- 4 · Zoom out: the whole dayOn days with AI use, the same people made about 200 more switches. Whether AI calms work or busy moments bring AI: unknown.
Source: Sumer S. Vaid, Ashley V. Whillans (Harvard Business School), “Digital Fragmentation and Generative AI Use Across 103 Million Application Events”, arXiv:2607.06681v2, preprint under review. Blocks are schematic, not to scale.
Graphic: Powered by Thorsten Meyer AI
Why AI Is Not a Focus Measure
The results challenge a simple claim that using an AI assistant automatically reduces digital distraction. The apparent decrease in switching immediately after AI use can coexist with more switching across an AI-use day. Those are different measurements, and relying on only one could produce an incomplete picture of how AI fits into work.
For employers considering AI tools as a way to improve concentration, the paper offers no test of productivity, work quality or employee well-being. It tracks application activity, not whether workers completed tasks faster or better. The authors also report that day-to-day variation within a single person accounted for 44.6% of differences in fragmentation, compared with 35.8% between people and 19.6% between organizations. That suggests the circumstances of a particular day may matter more than a stable label such as “distracted worker.”
The findings may support questions about meeting load, communication patterns and uninterrupted work periods, but they do not establish that any particular scheduling change will help. Nor do they show that AI tools are ineffective. They limit what can responsibly be claimed from this dataset.
Popular right now

Breville Barista Express Impress Espresso Machine with Grinder, Brushed Stainless Steel

Shark Rocket Pet Plus Corded Stick Vacuum, Bordeaux/Silver, HV322

Breville Barista Express Espresso Machine, Black Truffle
As an affiliate, we earn on qualifying purchases.
What the Tracking Captured
The paper, Digital Fragmentation and Generative AI Use Across 103 Million Application Events, was posted on arXiv on July 7, 2026, and revised on August 7. It identifies Vaid with Harvard Business School’s Artificial Intelligence Institute and Whillans with the school’s Negotiation, Organizations and Markets Unit. The manuscript says it is a preprint under review and that its findings could change during peer review.
The sample covered software, accounting, financial services, e-publishing, law and marketing organizations. A median of 58 workdays was recorded for each person. The study counted computer application activity, not phone use or all forms of work. It also compared the app-switching pattern with earlier attention research, but the authors caution that differences in methods make those comparisons a broad trend signal rather than a precise measure of change over time.
The headline that circulated on November 18 emphasized the reported average of roughly 900 daily switches and the short-term post-AI decline. The paper’s day-level result complicates that framing: AI-use days showed more switching overall. Both observations come from the same observational dataset, and neither establishes what caused the pattern.
“The manuscript is currently under review, and the findings may change based on the peer review process.”
— The paper’s authors
What App Logs Cannot Establish
The study is observational and cannot determine whether AI use leads to more fragmented days, whether workers turn to AI on days that are already fragmented, or whether both occur. It does not assess whether the work was completed more effectively, whether workers found the interruptions disruptive, or whether the recorded app activity represented meaningful task changes.
The sample is limited to eight organizations in India and computer activity recorded through Flowace. The findings do not establish whether the same patterns apply in other countries, workplaces, or on phones. The preprint is still under peer review, and its authors say the results may change. Employers also chose which applications were classified as productive, a measurement choice that affects how activity is interpreted.
Peer Review and Broader Tests
The immediate next milestone is peer review of the preprint. The researchers have not established a cause-and-effect result, so stronger evidence would require designs that can separate AI’s effects from the kinds of work and workdays on which employees choose to use it. Further studies would also need to test whether app-switching patterns correspond to output quality, task completion or workers’ own reports of focus.
For now, the paper supports a cautious reading: the post-AI sequence and the full-day pattern point in different directions. Employers should not treat the findings as proof that AI is a focus tool or a source of distraction. Whether the same associations hold across a wider range of workers and work settings remains untested.
Key Questions
Did the study find that AI reduces app switching?
It found fewer switches in the ten app uses after an AI use, but more switches across days when workers used AI. The result depends on the time window measured.
Did AI cause workers to switch apps more often?
The study cannot answer that. It is observational, so it shows an association and does not establish whether AI use caused the higher day-level switching or followed an already fragmented day.
How many workers and companies were included?
The researchers analyzed computer activity from 1,017 employees at eight organizations in India, covering several industries from November 2025 to January 2026.
Is the paper peer-reviewed?
No. The paper describes itself as a preprint under review and warns that its findings may change during peer review.
Does the study show AI improves productivity?
No. It measured application activity and switching, not work quality, task completion or productivity.
Source: ThorstenMeyerAI.com
