If you’re drowning in AI confusion, you’re not alone.
Yes, AI is transforming—and sometimes eliminating—routine tasks like data entry and basic processing. (Those jobs are toast.) But here’s what the hype machine misses: The revolution for professionals and knowledge workers isn’t about replacing ourselves with robots. It’s about something much more down-to-earth.
Effectively implemented, AI can free up strategic time for things that matter and that AI doesn’t do – those business essentials, such as learning, connecting with people, driving trust and intimacy, creative thinking, decision-making, and the actual execution of most aspects of our work. Oh, and occasionally seeing our families.
After months of trial and error with my own team, deep dives into research (with some help from Claude.ai), and extensive work with clients who have actual jobs (not just LinkedIn thought leadership positions), I’ve reached a modest yet profound conclusion.
I believe that the new goal for AI adoption — and the achievable goal among professionals and knowledge workers — is 20% time savings, not the 80% that is often dangled in front of us.
(Maybe this evolves to 40% someday, but we need to walk before we run.)
If I can do it, most professionals in thinking jobs—roles that require collaboration, analysis, and leadership—can too.
The Fallacy of 80% Productivity Gains (or Even 60%)
During my research into AI’s impact in the workplace, I kept stumbling across claims promising productivity gains approaching 80%. These numbers might seem reasonable—that is, if you work in a magical universe where meetings don’t exist, colleagues communicate telepathically, employees never get distracted, and humans run on photosynthesis instead of coffee breaks. But back here on Earth, the story is quite different.
Sure, studies from organizations like Nielsen Norman Group (claiming 66% gains) and Visual Capitalist (claiming 60% gains) look impressive on paper, but they’re based on inconvenient limitations: Typically, they are the result of controlled experiments that look at practices in isolation, not in the chaotic reality of actual workplace implementation. Even more inflated claims come from the marketing and sales hype (or wish-fulfillment) of service providers, consulting firms, and evangelists.
Here’s the thing: AI often adds time to professional workloads rather than trimming them. Digging through countless studies (seriously, a rabbit hole worthy of Alice herself), I couldn’t find a single credible source supporting 80% productivity gains in real-world environments. I even found studies claiming AI increases the time for workers. The AI evaluation company METR recently released a study that found developers took “19% longer” when using AI tools. An Upwork survey found “almost 80% of workers who use generative AI said it has added to their workload”.
The evidence shows a more nuanced (and slightly humbling) reality. Instead of delivering the massive time savings being promised, AI implementation often introduces new layers of complexity, demands substantial training time, and requires additional review time from professionals who are now essentially babysitting their digital assistants.
The 80% productivity gain narrative isn’t just overly optimistic, it’s completely disconnected from how AI functions for professionals “in the wild.”
How to Save 20% of Your Time
The answer, though, is not to throw up your hands and wait for someone else (or your corporate bosses) to figure it all out. The gains of AI can be exponential when the right goals and scope are married to the right capabilities. The 20% AI goal guides us to leverage AI where we will get the biggest gains without sacrificing our thinking, creativity, and personal skills.
No two professionals will save 20% of their time in the same way. The best approach? Take a hard look at what you spend your day doing. In my case, some of my work involves routine business operations, but I also handle completely unpredictable project work.
Although I’m an accountant, I no longer crunch numbers all day. But I have clients using AI tools like Copilot who are saving serious time on Excel formulas, numerical analysis, and data work. The key is finding your sweet spots.
For project work that is in-between “rote” and “novel,” like research tasks, reports, proposals, and so much more, AI platforms like Claude.ai can be trained to follow the same steps you do, and can use proprietary source material you provide, thus improving the quality of the AI output and saving time with the “babysitting” aspect of reviewing the deliverables.
Here’s what I mean: the point of time-savings isn’t to hand over everything to AI, especially areas like note-taking, writing, and analysis. That actually undermines how we learn best.
- Take note-taking, for example. The real advantage of good old-fashioned note-taking has nothing to do with the notes themselves and everything to do with how they support our thinking. Note-taking forces you to process information, figure out what matters from what doesn’t, and synthesize what you’ve collected. Studies consistently show that manual note-taking boosts memory and recall way better than passive consumption.
- Writing works the same way—it activates how we naturally learn best. Your brain forms better connections through the “tedious” effort of spelling out your thoughts. Writing isn’t just about the final product, it actually solidifies thinking and speeds up learning. I’m literally using writing to clarify my thoughts right now as I work on this article!
- Analyzing and thinking need to be protected. In my world of consulting, internal controls, and internal audit, AI excels at generating ideas, planning, cleaning up drafts, and certain types of testing. But having it do all your thinking and analysis? That’s not just undesirable—it’s actually dangerous. If you let AI handle all your emails, writing, and reports, you’re potentially stunting your intellectual growth. That’s not exactly the outcome we’re going for.
The Math of 20% AI Productivity Goals
The 20% may feel underwhelming compared to many claims. But let’s do the math. If you work 8 hours a day, 20% savings equals 1.6 hours daily. That’s huge—time for learning, actually connecting with colleagues, or getting home for dinner. Over a 40-hour week, that’s 8 hours back in your life. Over a working year? We’re talking almost 400 hours!
Now, let me set expectations here. The 20% goal is my across-the-board target, but it’s not uniform. Some tasks become 50% faster. Others improve by 40%. Sometimes it’s only 5%. And when you’re learning? You may lose time experimenting with AI. I recently spent over an hour trying to get various tools to format something into a table. It didn’t work (this time), so I went back to doing it manually. There’s always next time.
To leverage your personal 20% sweet spot, ask yourself:
- What tasks eat up most of your time?
- What projects are currently on your plate?
- How can you invest some time thinking strategically about these opportunities?
Where Does This 20% Goal Take Us?
Let’s be real for a minute. The economy has been taking a beating lately, at least here in Calgary. I talk to companies daily that have hiring freezes and consultant moratoriums. You could argue that finding 20% efficiency isn’t just a competitive advantage—it’s the competitive reality you need to stay afloat. That’s sobering, maybe a bit depressing, but unfortunately pretty realistic.
But here’s where I get genuinely excited about this shift. When I look at knowledge work overall, I see us moving away from low-value, mind-numbing rote work—and that’s actually fantastic. In my world of internal audit and internal controls, I’ve watched this transformation unfold over the past decade. We’ve moved away from manual, transactional testing and template-filling toward value-added advisory services.
Generative AI is turbocharging this movement by challenging the complacency that creeps into our professions and corporate worlds. It’s especially powerful for larger enterprises that can leverage these programs at scale and that have more waste to find than their smaller competitors.
There is also an AI Learning Paradox. I see with my more enthusiastic AI friends and coworkers: that the time we save often goes right back into learning more about AI! Will this become a vicious or virtuous circle? The jury’s still out.
You can’t just download ChatGPT and expect miracles. You need to take an honest look at your current work patterns, experiment with tools systematically, and develop new habits that stick.
Whether you’re looking to implement time-saving AI strategies for yourself, get your entire team up to speed through workshops, or identify broader process improvements that could benefit from intelligent automation, the key is starting with a realistic plan grounded in real-world experience.
The 20% is there for the taking—but only if you’re willing to do the work to get it.
Want Help with Real-World Guidance to Reclaim Your 20%?
If you’d like help implementing these time-saving strategies in your organization or want to explore how AI tools can streamline your specific workflows, let’s talk. Risk Oversight works with professionals and teams to develop practical AI adoption strategies, facilitate hands-on workshops, and identify process improvements that deliver measurable outcomes. Contact me at adrienne@riskoversight.ca to discuss how we can help you and your team move from AI confusion to AI productivity—without the hype, just real solutions.