Employees at companies actually using AI are saving 40–60 minutes every single day. For a 50-person team, that's 33–50 hours of extra capacity — daily. And yet 81% of U.S. businesses still aren't using AI at all.
What Is It?
Goldman Sachs economists Sarah Dong and Joseph Briggs released an AI Adoption Tracker in March 2026. It's based on data from the U.S. Census Bureau's Business Trends and Outlook Survey — and the findings are pretty striking.
The key is that companies using AI are clearly pulling ahead — while those that aren't haven't even entered the race.
The headline number: employees at companies using ChatGPT Enterprise save an average of 40–60 minutes per day, and 75% say they're now completing tasks they simply couldn't do before.
Academic studies report an average 23% productivity gain, while real-world business cases show around 33% efficiency improvement. Anthropic's own analysis estimates that current-generation AI models could lift U.S. labor productivity growth by 1.8% annually.
What Changes?
Here's the thing — this gap isn't closing. It's widening.
| AI Adopters (19%) | Non-Adopters (81%) | |
|---|---|---|
| Daily time saved | 40–60 min/person | 0 minutes |
| Productivity gain | 23–33% | No change |
| Product launch cycle | 24–36 months → 6 months | Same as before |
| Employee capability | 75% completing new types of work | Skills unchanged |
The gap is just as clear when you break it down by company size:
Industry divides are just as stark. Computing and web hosting are already at 60% adoption, and broadcasting is projected to see the largest uptick over the next six months. That's a clear signal that AI adoption in media and content is about to accelerate.
And once people start using AI, usage compounds. According to OpenAI, message volume from business users jumped 30% in just a few months. The more you use it, the more you use it.
So Why Are 81% Still Sitting It Out?
Here's where it gets interesting. 77% of companies say they're actively pursuing AI — but only 19% are actually using it. There's a massive gap between "we plan to" and "we are."
77% say they plan to adopt AI vs 19% who actually have — that 58-point execution gap means most companies simply don't know how to evaluate, buy, and deploy AI.
Pull together the research from Deloitte, Gartner, and Bain, and the barriers come down to three things:
- Skill gaps among employees
According to Deloitte, employee skills are the single biggest barrier to integrating AI into existing workflows. Buying the tools doesn't help if no one knows how to use them. - Data security concerns
Can we feed company data into AI? Most organizations can't give a clear answer to that question. A Cloudera survey found 80% of companies say data access issues are blocking their AI adoption. - Difficulty measuring ROI
44% of CFOs say they plan to cut headcount using AI, yet they also admit that measured productivity gains haven't matched expectations. Proving ROI in concrete numbers is still genuinely hard.
But those barriers are clearly coming down. Bain found that in more than 80% of companies that have adopted AI, results met or exceeded expectations. The case for skepticism is getting weaker.
Getting Started
The Goldman Sachs data makes one thing clear: the risk of waiting is greater than the risk of starting. If a 50-person team is missing out on 33–50 hours of productivity every day, that's essentially donating that time to your competitors.
- Pick one repetitive task
Email drafts, data cleanup, research summaries — find one task that eats 30+ minutes a day and hand it to AI. Don't try to roll it out company-wide. Start with one task. - Define the problem before picking a tool
Choosing the tool first is a recipe for failure. Define the problem you want to solve, then find the right tool for it. The reason 77% of companies stall on adoption is they buy tools before defining the problem. - Run a 2-week pilot and measure it
Have a team of five use it for two weeks and track the time saved. What your CFO needs isn't "AI is great" — it's "we saved X hours per week." - Set security guidelines upfront
A one-page guide on what data can and can't go into AI is enough. Let's be honest — banning AI because there's no policy in place is worse than having a simple one. - Share small wins internally
As OpenAI's data shows, AI usage compounds. When one person shares a success story, the whole team starts to follow.
Key Takeaway: If your competitors gain an extra hour of productive work every single day, that's a 130-hour advantage after six months. In a world where product cycles are moving 4x faster, "let's wait and see" isn't a strategy — it's closer to giving up.
Deep Dive Resources
Goldman Sachs AI Adoption Tracker
Monthly tracking of AI adoption rates by industry and company size, based on Census Bureau data. The primary source for the numbers in this post.
Anthropic — AI Productivity Estimation Research
Analysis of 100,000 Claude conversations estimating that AI speeds up individual tasks by 80% and could lift U.S. labor productivity growth by 1.8% annually.
Bain — AI Moves from Pilots to Production
Among the 59% of companies that have adopted AI, 80% report results meeting or exceeding expectations. Real-world data on the critical transition from pilot to production.
