Voice AI and AI agents at work: 2026 statistics

22 figures on how work is changing, from app-switching and AI-agent adoption to voice and measured time saved. Every number links to its primary source.

Last updated July 2026 · every figure is quoted from a named source, with the year shown

In short

The modern workday is scattered across dozens of tools. The average worker toggled between apps and websites about 1,200 times a day (Harvard Business Review, 2022) and is interrupted roughly every two minutes (Microsoft, 2025). Most of that switching is the busywork a cross-app assistant is meant to absorb.
AI agents are moving from experiment to default. Two research firms put the AI agents market near $50 billion by 2030, growing about 46% a year (MarketsandMarkets and Grand View Research, 2025), and Gartner predicts 40% of enterprise apps will include task-specific agents by the end of 2026.
Most companies use AI, but few see the payoff yet: 88% of organizations use AI in at least one business function while only 39% report companywide impact (McKinsey, 2025). The gap is between having AI and having it do real work.
Trust, not autonomy, is the constraint. Gartner expects more than 40% of agentic AI projects to be canceled by 2027 (2025), and trust in fully autonomous agents fell from 43% to 27% in a year (Capgemini, 2025). The agents people keep are the ones that ask before acting.
Voice is a faster way in. A Stanford study found speaking text is 3.0 times faster than typing it on a phone (2016), and US voice-assistant users are forecast to reach 168 million by 2029 (eMarketer, 2025).

The problem

How much time do we lose switching between apps?

Work is spread across more tools than ever, and the cost is the constant switching between them. The numbers below are why an assistant that acts across your apps from one request, instead of making you click between them, saves real time.

1,200The average worker toggled between apps and websites about 1,200 times a day.Harvard Business Review, 2022
~4 hrsThat toggling added up to just under four hours a week reorienting, about five working weeks a year.Harvard Business Review, 2022
2 minPeople using Microsoft 365 were interrupted every two minutes by a meeting, email, or ping.Microsoft Work Trend Index, 2025
9.5 minIt took workers 9.5 minutes on average to get back into a good workflow after switching apps.Qatalog and Cornell University, 2021
58%Knowledge workers spent 58% of the day on repetitive coordination, or “work about work”.Asana Anatomy of Work, 2023
93The average company ran 93 different software apps, up 4% from the year before.Okta Businesses at Work, 2024

AI agents at work

How fast are AI agents being adopted at work?

AI agents, software that takes a goal and carries out the steps, are moving from pilots to production. Analysts broadly agree on the direction, and they also agree on the catch: the projects that stick are the ones people can trust.

~$50BTwo firms put the AI agents market near $50 billion by 2030, growing about 46% a year.MarketsandMarkets and Grand View Research, 2025
40%Gartner predicts 40% of enterprise apps will include task-specific AI agents by the end of 2026, up from under 5% in 2025.Gartner, 2025
15%Gartner predicts 15% of day-to-day work decisions will be made autonomously by AI agents by 2028, from 0% in 2024.Gartner, 2025
62%62% of organizations were at least experimenting with AI agents, but no more than 10% were scaling them in any one function.McKinsey, 2025
40%+Gartner predicts more than 40% of agentic AI projects will be canceled by the end of 2027, citing cost and unclear value.Gartner, 2025
43% to 27%Trust in fully autonomous AI agents fell from 43% to 27% in a single year.Capgemini Research Institute, 2025

Voice at work

Is voice a real way to get work done?

Voice is not just for phones and speakers anymore. Speaking is simply faster than typing, and the tools that understand it are getting better and more widely used every year.

3xA Stanford study found speaking text was 3.0 times faster than typing it on a phone keyboard.Stanford and University of Washington, 2016
168MUS voice-assistant users are forecast to grow from 139.8 million in 2022 to 168.2 million by 2029.eMarketer, 2025
$23BThe speech and voice recognition market is projected to grow from $9.66 billion in 2025 to $23.11 billion by 2030.MarketsandMarkets, 2025
$80BGartner predicted conversational AI would cut contact-center labor costs by $80 billion in 2026.Gartner, 2022

AI at work

How many people use AI at work, and does it actually save time?

Adoption is nearly everywhere; measured payoff is more modest and uneven. The most rigorous studies find real gains, largest for less-experienced workers, which is the honest picture behind the headlines.

51%51% of US employees used AI at work at least occasionally by late 2025, and 12% used it daily.Gallup, 2026
37%The share of US workers using generative AI on the job rose from 28% in 2024 to 37% in 2025.Bick, Blandin and Deming (NBER), 2025
88% vs 39%88% of organizations used AI in at least one business function, but only 39% reported companywide bottom-line impact.McKinsey, 2025
14%In a controlled study, AI support raised issues resolved per hour by 14% on average, and 34% for the least experienced workers.Brynjolfsson, Li and Raymond (NBER), 2023
~30 minIn a randomized trial, workers who used an AI assistant regularly spent about 30 minutes less on email each week.Microsoft Research, 2025
60-70%Generative AI and other technologies could automate work activities that absorb 60 to 70% of employees’ time today.McKinsey, 2023

How we compiled this

Every figure on this page was taken from a named primary source, 19 in all, across analyst firms, academic studies, and first-party surveys, and checked against the source’s own wording. Figures published between 2016 and 2026 are shown with the year they were reported, and older ones are framed as forecasts or as of their date. Where research firms disagreed, we show the range rather than the biggest number. We update the page when the underlying figures change.

Sources

Spend less of the day switching apps.

Incredible acts across your apps from one spoken request. Download it and hand the busywork over.

Download See how it compares

Statistics are the property of their respective sources, linked above, and are cited here for reference. Read more about agentic AI or the terms around Incredible.