A Harvard Study Examines AI As Workers Bounce Between Apps
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: A Harvard Study Examines AI As Workers Bounce Between Apps on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get the latest gadgets delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

Harvard researchers analyzed 103 million application events from 1,017 office workers in India. They found fewer switches in the ten app uses after AI use, but about 200 more switches on days workers used AI; the observational preprint cannot establish why.

A Harvard Business School preprint analyzing 103 million application events from 1,017 office workers found that app switching was lower in the ten app uses after workers used AI, but higher on days when they used AI at all. The mixed results do not show that AI causes either pattern, and the study remains under peer review.

Researchers Sumer S. Vaid and Ashley V. Whillans examined activity recorded from November 2025 through January 2026 at eight companies in India. The sample included employees in software, accounting, financial services, e-publishing, law and marketing. The paper says the researchers analyzed anonymized records gathered through Flowace, a commercial workforce-analytics service already used by the participating organizations.

Across 62,693 person-days, workers switched between applications an average of 908.3 times per day; the median was 814. These counts cover changes to a different application, not every change of document or browser tab. The paper also reports that a worker stayed in an application for an average of 27.8 seconds before switching, with a median of 3.3 seconds. Those figures describe the study’s specific measures and sample, not all workers or all forms of digital activity.

The AI findings point in two directions. In the ten application uses following an AI use, people used fewer apps, switched less, stayed about half a second longer per app and followed more predictable sequences. Yet on days with AI use, workers made about 200 more switches per day, and days with more AI minutes were more fragmented. The paper does not establish whether AI use led to more fragmented days, whether workers used AI on days that were already more fragmented, or whether both patterns reflect other factors.

At a glance
reportWhen: Preprint posted July 7, 2026, revised A…
The developmentA Harvard Business School preprint reports mixed associations between generative AI use and app switching in workplace data from eight Indian companies.

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

Why AI’s Daily Pattern Matters

The findings caution against treating an AI tool as a proven way to reduce workplace distraction. A lower switching rate immediately after AI use can coexist with more switching across the day, so a narrow measure of activity after a prompt may not describe an employee’s overall workday. The study measures application behavior, not work quality, output, stress or productivity.

For employers considering AI products or monitoring practices, the distinction matters: a count of app changes cannot by itself show whether workers are getting more done or concentrating better. The authors’ analysis also suggests that day-to-day conditions may matter more to fragmentation than stable differences between workers or organizations. That points to work scheduling and communication demands as issues worth examining, rather than assuming that buying an AI tool will solve interruptions.

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 Workplace Data Was Collected

The paper, titled Digital Fragmentation and Generative AI Use Across 103 Million Application Events, was posted to arXiv on July 7, 2026, and revised on August 7. Its PDF labels it Preprint version 4.0 and says the manuscript is under review, with findings subject to change. It is not yet a peer-reviewed journal publication.

Flowace software recorded when the active application changed, along with information such as the application name and session times. Participating employers determined which apps counted as productive; the researchers received anonymized data under a data-use agreement. The paper says Harvard’s human-subjects committee approved the protocol. It also says the research had no external funding and Flowace had no role in study design, analysis or interpretation. One author disclosed equity in a private company unrelated to Flowace.

The authors report that 44.6% of variation in fragmentation occurred across different days for the same person, compared with 35.8% between people and 19.6% between organizations. They also report that fragmentation peaked on Tuesdays and reset after weekends and holidays. These are findings within this sample; the study did not include workers in other countries or activity on phones.

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

— The authors, in the paper’s preprint notice

What the Study Cannot Establish

The research is observational, not experimental. It cannot determine whether AI use changes app-switching behavior or whether people turn to AI when their workday is already unusually fragmented or consolidated. The authors themselves describe these possibilities and do not resolve the direction of cause and effect.

The sample is limited to 1,017 employees at eight Indian organizations whose employers used one workforce-monitoring service. It does not establish whether the same patterns apply in other countries, industries, workplaces without this software, or on phones. The paper measures app events, not whether workers completed tasks faster, produced better work or experienced more distraction. Its results could also change through peer review.

Peer Review and Broader Evidence

The immediate next milestone is peer review; the paper’s authors warn that its findings may change. Further studies would be needed to test whether the patterns hold across different countries and work settings and to examine whether AI use causes changes in switching behavior. Research linking app activity to work quality or other outcomes would also help clarify what the measured fragmentation means for employees.

Until then, the study offers a detailed account of app activity in its particular sample, not a general verdict on AI and focus. Employers and workers can treat its results as a reason to examine how workload and communication vary during the week, but the paper does not support buying AI tools on the premise that they will reduce app switching.

Key Questions

Did the Harvard study find that AI reduces app switching?

Not across the whole day. Switching was lower in the ten app uses after an AI use, but workers made about 200 more switches on days when they used AI. The study cannot establish why.

How many workers and companies were included?

The researchers analyzed records from 1,017 office workers at eight companies in India, covering November 2025 through January 2026.

Is the paper peer-reviewed?

No. It is a preprint under peer review. The paper says its findings may change during that process.

Does the study show that AI improves productivity?

No. It measures application events and switching patterns, not productivity, work quality or task completion.

Source: ThorstenMeyerAI.com

You May Also Like

Apple rejected my dictation app for using the accessibility API

A developer’s Mac dictation app, WhisperPad, was rejected by Apple for using the accessibility API, raising questions about app guidelines and accessibility tools.

Show HN: Whiteboard (YC W26) – An Open-source IDE For Thoughtful Software Design

Whiteboard is an open-source desktop IDE aimed at fostering deliberate software design, developed by YC W26 founders. Its release sparks interest in thoughtful coding tools.

Disk Is the Contract: Inside Threlmark’s Local-First Architecture

Threlmark innovates with a local-first, file-based architecture, making project data portable, inspectable, and resilient without a database.

The Emacsification of Software

AI-driven customization transforms software, echoing Emacs’ flexible culture, leading to more personalized, native interfaces and tools.