A one-minute estimate of how much time spent drafting documents and reports, handling customer emails or extracting data from invoices generative AI could free up. Coefficients based on public studies from McKinsey, MIT, Stanford, GitHub and Gartner: the formula is visible and you can adjust every assumption.
Think of drafting documents and reports, answering customer emails and queries, extracting data from invoices and documents, or searching and summarizing internal information. According to McKinsey, generative AI combined with other technologies can automate activities that today absorb 60% to 70% of working time.
Employer cost per worked hour (salary + social contributions) of administrative staff.
Each selected task adds 20% coverage over total AI-automatable time. It is a configurable split, not a measurement: adjust the selection to your case.
Conservative (15%): in line with the +14% productivity gain measured in customer support (Stanford/MIT). Medium (25%): midpoint between that result and the savings in professional writing. Ambitious (40%): matching the -40% time reduction in the MIT writing experiment published in Science; in coding, GitHub measured up to 55%.
We do not make the percentages up. The scenarios in this calculator are based on studies published by leading institutions and companies. Here they are, with their original links.
According to McKinsey, generative AI could add between $2.6 and $4.4 trillion annually to the global economy across the 63 use cases analyzed, and combining generative AI with other existing technologies has the potential to automate activities that today absorb 60% to 70% of employees' time.
McKinsey & Company — The economic potential of generative AI (2023) →A study by Stanford and MIT economists covering 5,179 customer support agents showed that a generative AI conversational assistant increased productivity —measured in issues resolved per hour— by 14% on average, and by up to 34% for novice or lower-skilled agents, while also improving customer satisfaction and reducing attrition.
NBER (Stanford/MIT) — Brynjolfsson, Li y Raymond, "Generative AI at Work" (2023) →In a randomized experiment published in Science by MIT researchers, 453 college-educated professionals who used ChatGPT for professional writing tasks reduced completion time by 40% on average and increased output quality by 18%, with the largest gains among workers with lower initial skill.
Science (MIT) — Noy y Zhang, "Experimental evidence on the productivity effects of generative AI" (2023) →In a controlled GitHub experiment with 95 professional developers, those using GitHub Copilot completed a programming task (an HTTP server in JavaScript) 55% faster than the group without the tool: 1 hour 11 minutes versus 2 hours 41 minutes on average (statistically significant, P=0.0017).
GitHub — Quantifying GitHub Copilot's impact on developer productivity and happiness (2022) →In a randomized controlled trial by GitHub with Accenture developers in a real enterprise environment, using GitHub Copilot increased pull requests per developer by 8.7%, merge rate by 15% and successful builds by 84%; developers accepted around 30% of suggestions and 90% felt more fulfilled with their job.
GitHub y Accenture — Quantifying GitHub Copilot's impact in the enterprise (2024) →According to Gartner, avoidable rework caused by human error can consume up to 30% of an accounting employee's time, and robotic process automation (RPA) in financial reporting can save finance departments 25,000 hours of avoidable work per year —worth about $878,000— in an organization with 40 full-time accounting staff.
Gartner — RPA can save finance departments 25,000 hours of avoidable work annually (2019) →According to a Federal Reserve Bank of St. Louis survey (Real-Time Population Survey, November 2024), workers who use generative AI save on average 5.4% of their weekly work hours, equivalent to about 2.2 hours per week on a 40-hour schedule; among frequent users, one in five reported saving 4 or more hours per week.
Federal Reserve Bank of St. Louis — "The Impact of Generative AI on Work Productivity" (2025) →In Thomson Reuters' "Future of Professionals" survey of more than 2,200 professionals in legal, tax & accounting, and risk & compliance, respondents expect AI to save them 4 hours of work per week over the next year —about 200 hours annually— and up to 12 hours per week by 2029.
Thomson Reuters — Future of Professionals Report 2024 →The number above is a starting point. With myIA we identify your specific AI use cases, prioritize them by ROI and implement them with real adoption by your team.