What AI Can—and Can't—Do About App Switching, According To Harvard
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: What AI Can—and Can't—Do About App Switching, According To Harvard on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

  • Business-only prices and quantity discounts
  • Tax-exempt purchasing
  • Multiple users, one account, clear invoices
As an affiliate, we earn on qualifying purchases.

TL;DR

A Harvard Business School preprint analyzed 103 million application events from 1,017 workers at eight organizations in India. Workers switched apps less in the ten app uses after an AI use, but made about 200 more switches on days they used AI; the observational data cannot show whether AI caused either pattern.

A Harvard Business School preprint analyzing 103 million application events from 1,017 office workers found a mixed relationship between generative AI use and app switching: employees switched among fewer apps in the ten uses after an AI interaction, but made about 200 more switches on AI-use days than on days without recorded AI use. The study is observational, so it does not establish that AI caused either pattern or that it improved work.

The paper, “Digital Fragmentation and Generative AI Use Across 103 Million Application Events,” was written by Sumer S. Vaid and Ashley V. Whillans of Harvard Business School. It analyzes 56.9 million app switches across 62,693 person-days, using records collected from November 2025 through January 2026 at eight organizations in India. The manuscript identifies itself as a preprint under peer review and cautions that its findings may change.

Across the full sample, employees switched applications a mean of 908.3 times per workday; the median was 814. The middle half of employees recorded between 547 and 1,130 switches per day. An application remained active for a mean of 27.8 seconds before a switch, while the median was 3.3 seconds. These measures count changes between applications, not movement between documents inside the same app.

The AI findings vary by time scale. In the ten application uses following an AI use, workers used fewer distinct apps, switched less often, stayed roughly half a second longer per app and moved through apps in a more predictable pattern. Across whole days, however, recorded AI use coincided with roughly 200 additional app switches, and days with more AI-use minutes were more fragmented. The source material does not provide a causal estimate showing what would have happened to the same worker’s day without AI.

At a glance
reportWhen: Preprint posted July 7, 2026, revised A…
The developmentA Harvard Business School preprint reports that AI use was associated with fewer app switches immediately afterward but more switching across AI-use days overall.

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.

10 app uses before AI median 7 seconds 10 app uses after Whole AI day: about 200 more switches
  1. 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. 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. 3 · The next ten app usesFewer different apps, fewer switches, about half a second longer per app, more predictable. Real, but small.
  4. 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

AI’s Effects Differ Across the Workday

The results challenge a simple claim that adding AI automatically reduces digital interruptions. A brief stretch of more orderly app use after an AI interaction can coexist with a more fragmented day overall. That distinction matters to employers considering AI software as a way to improve focus: this study does not show that AI is a focus tool, nor does it establish that switching more means workers are less productive.

The authors also report that fragmentation varied more from one day to another for the same person than between organizations: 44.6% of variation was attributed to within-person day-to-day changes, compared with 35.8% between people and 19.6% between organizations. This suggests that workload and daily conditions may be relevant to how app switching unfolds. It does not identify which conditions drive those changes, or whether reducing switches would improve output, wellbeing or work quality.

Products Worth Considering

DJI Mini 4K Drone Bundle with Strobe Light, Landing Gear, and Propeller Holder Strap

DJI Mini 4K Drone Bundle with Strobe Light, Landing Gear, and Propeller Holder Strap

As an affiliate, we earn on qualifying purchases.

How the App Records Were Collected

The researchers used data from Flowace, a commercial workforce-analytics service already used by the participating organizations. According to the paper, its software records an event when an employee changes the active application, including the app name and timestamps. The employers, rather than the researchers, determined which apps were classified as productive. The participating organizations worked in software, accounting, financial services, e-publishing, law and marketing.

The study’s authors say the data were anonymized before researchers received them, access was governed by a data-use agreement, and Harvard’s human-subjects committee approved the protocol. The paper states it received no external funding and says Flowace had no role in study design, analysis or interpretation. One author disclosed equity in an unrelated company that uses AI to create enterprise digital twins. The sample covers work computers at eight Indian organizations; it does not establish whether the patterns apply to other countries, phone use or different kinds of workplaces.

The paper was first posted to arXiv on July 7, 2026, and revised on August 7. Its authors explicitly describe it as under review. The headline shared on November 18 emphasized reduced switching after AI use, but the same paper reports more switching on AI-use days—two observations that refer to different time windows and should not be treated as contradictory proof of a single effect.

“The manuscript is currently under review, and the findings may change based on the peer review process.”

— Vaid and Whillans, in the preprint

Causation and Work Quality Remain Unknown

The study tracks app activity rather than assigning workers to use or avoid AI. It therefore cannot determine whether AI led to extra switching, whether workers chose AI on days that were already more fragmented, or whether both patterns reflected other demands. The authors say the direction of the relationship is unknown.

The analysis also does not establish whether frequent switching harmed concentration or reduced output. The available source material does not report measures of work quality, completed tasks or employee wellbeing. Results from eight organizations in India may not generalize to other countries or industries, and the paper’s peer-review status means its methods or conclusions could change.

Peer Review and Broader Evidence

The immediate next milestone is the paper’s peer-review process; the authors warn that findings may change before publication. Further research would need to test whether AI use causes changes in switching, account for what work people are doing when they use AI, and measure outcomes beyond app activity. Until then, the findings describe associations in one set of monitored workplaces, not a general rule about AI and focus.

For readers and employers, the paper supports caution when interpreting claims that AI reduces workplace interruptions. Its headline numbers describe recorded app use during the study period; they do not show that buying an AI tool will make employees more focused or productive.

Key Questions

Did the Harvard study find that AI reduces app switching?

Not across the full day. Workers switched among fewer apps in the ten app uses after an AI interaction, but made about 200 more switches on days when they used AI.

How many workers and app events were analyzed?

The preprint analyzed 103 million application events from 1,017 workers at eight organizations in India, covering November 2025 through January 2026.

Does the study prove AI causes more switching?

No. It is an observational analysis, not an experiment. The authors say it is unclear whether AI use changes switching, workers use AI on already fragmented days, or both.

Is the research peer-reviewed?

The paper was posted on arXiv and describes itself as under peer review. Its authors say the findings may change during that process.

Does the study show that workers became more productive?

No. The reported analysis measures application use and switching. The source material does not establish whether work quality, output or employee wellbeing improved or declined.

Source: ThorstenMeyerAI.com

You May Also Like

Anthropic’s Internal AI Bill Tops Employee Pay At Customer Rates

Business Insider reports Anthropic’s internal AI use exceeds employee pay at customer prices, but the calculation and scope are unavailable.

SenseTime-W Announces Profitable Interim Results Amid AI Revenue Surge

SenseTime-W announces RMB 607M profit and 28.2% growth in generative AI revenue amid sector-wide competition and strategic shift.

What Meta And Microsoft’s Claude Cutback Means For Internal AI Use

A Yahoo Finance headline says Meta and Microsoft are reducing internal Claude use, but the available information gives no scope, timing or reason.

AI In Government: Europe’s Jobs, Asylum Files And 90,835 Replies

Dutch job-cut plans, UK asylum summaries and bot responses to a health consultation raise questions about AI use and oversight in government.