The Hidden Cost of
Workplace Fragmentation

Gloria Mark, Attention Span: A Groundbreaking Way to Restore Balance, Happiness and Productivity (Hanover Square Press, 2023), and the UC Irvine studies behind it: average attention on a screen fell from about 2.5 minutes in 2004 to 47 seconds, and it takes about 23 minutes to return to the original task after an interruption. The 23-minute figure is hers; it is often quoted with no source at all.

Microsoft Work Trend Index, 2023 and 2024 editions: app and window switching (about 1,200 toggles a day), and the reorientation cost that accumulates to roughly five working weeks a year. microsoft.com/worklab/work-trend-index

Asana, Anatomy of Work Index: the $450 billion annual cost of context switching to the US economy. A 2021 to 2022 estimate, cited here as such rather than as a current-year figure.

Microsoft Work Trend Index: 48 percent of employees describe their work as chaotic and fragmented, with interruptions arriving about every two minutes during the working day.

McKinsey Global Institute, The Social Economy and successor work: knowledge workers spend about 28 percent of the week on email and about 39 percent on role-specific work, with the remainder on coordination.

Zylo and Spendesk SaaS management reports (2023): the average enterprise runs 473 SaaS applications; more than half of licences go unused for 90 days or more; organizations use about 47 percent of what they buy.

Microsoft Work Trend Index (2024): 78 percent of AI users bring their own AI tools to work, outside IT governance.

Microsoft 2026 Work Trend Index, Agents, human agency, and the opportunity for every organization (May 2026): active agents in the Microsoft 365 ecosystem grew 15 times year on year, and 18 times at large enterprises. microsoft.com/worklab

Sophie Leroy (University of Washington), research on attention residue: switching tasks leaves cognitive traces that measurably impair performance on the task switched to.

Enterprise communication cost studies: email and internal messaging estimated at $28,209 per employee per year, of which roughly $9,500 is attributed to chat. An industry estimate, not a peer-reviewed measurement, and treated as one.

SaaS sprawl and retention research: tool fragmentation produces data silos and duplicated analysis, and application fatigue is a documented contributor to attrition among high performers.

Verizon, 2026 Data Breach Investigations Report: breaches involving a third party rose from 30 percent to 48 percent in a year, across more than 31,000 incidents and 22,000 confirmed breaches (1 November 2024 to 31 October 2025). verizon.com/business/resources/reports/dbir

Verizon 2026 DBIR, shadow AI section, drawing on LayerX browser telemetry: 858,440 data-loss events aimed at generative AI tools; shadow AI is now the third most common non-malicious insider action, a fourfold rise year on year, with source code the leading data type sent to an external model.

LayerX, Enterprise AI and SaaS Data Security Report 2025: 77 percent of employees paste into GenAI prompts, 82 percent of those from unmanaged accounts; about 22 percent of pastes carry PII or payment data; about 40 percent of files uploaded to GenAI contain PII or payment data. layerxsecurity.com

LayerX, Browser Extension Security Report 2026: 20.63 percent of enterprise users have at least one AI-enabled browser extension installed, and 73 percent of AI extensions carry high or critical permission scope.

IBM, Cost of a Data Breach Report 2026 (29 July 2026): global average breach cost $4.99 million; one in four malicious breaches is AI-enabled, a 56 percent rise, at about $6 million; more than 20 percent of organizations reported a breach targeting an AI model or application; the leading causes were compromised APIs, applications or plug-ins (27 percent) and cloud misconfiguration of AI workloads (27 percent); only 37 percent of breached organizations encrypt sensitive data both at rest and in transit. newsroom.ibm.com

IBM, Cost of a Data Breach Report 2025: shadow AI added roughly $670,000 to the average breach; 63 percent of breached organizations had no AI governance policy. Retained as the 2025 figure it is, not carried forward into 2026.

Schellman, State of AI Governance 2026 (more than 500 US enterprise leaders, published 16 September 2026, summarized by the Cloud Security Alliance): 74 percent believe they could pass an AI compliance audit while 27 percent report a fully mature programme; 44 percent have documented AI incident response; 29 percent feel prepared for the EU AI Act; 36 percent of boards regularly discuss third-party AI risk. cloudsecurityalliance.org

European Commission, AI Act regulatory framework and implementation timetable: enforcement and transparency obligations from 2 August 2026; high-risk obligations for stand-alone Annex III systems deferred to 2 December 2027 and for AI embedded in regulated products to 2 August 2028. digital-strategy.ec.europa.eu

Regulation (EU) 2024/1689 (the AI Act), Article 99: penalties up to 35 million euros or 7 percent of worldwide annual turnover for prohibited practices, and up to 15 million euros or 3 percent for most other infringements.

Stanford HAI, AI Index 2026: the share of firms reporting no responsible-AI policy of any kind fell to about 11 percent, from about 24 percent a year earlier. Cited to make the opposite of the obvious point: policy coverage is rising, so the question moves from whether a policy exists to whether its operation can be evidenced.