There are two types of people in AI right now: 1. Those who never block time to AI-ify their work… 2. And those who’ll spend all day automating a 15-minute task. Both are leaking leverage. On one side: “I’ll systemize later.” (They never do.) On the other: 6 hours deep in Zapier trying to perfect something that saves 15 minutes. The leverage is in the middle. 😉 Take something simple like rescheduling meeting conflicts. It’s a 15-minute task. You could just do it. Or… Spend 60 minutes building a clean playbook so AI handles it next time. 15 minutes today + 60 minutes to systemize -> 2 minutes forever after (!!) But I wouldn't spend 6 hours on it. Here's what we're trying: 1. Choose a % of your time. Right now, I'm at 10% 2. Invest that, no questions asked, in AI-ifying your work 3. Yes you can do more, but you need more discernment for higher investments This keeps you from the "I'm too busy" excuse. But also keeps you from over-engineering it. The time invested in how you work is where real leverage compounds.
The problem is that as much as most people want to spend 60 mins creating a clean playbook and having a nice automated system in place, they don't know how. They're used to working with a tool like ChatGPT and left scratching their head at how to bring AI into their actual workflows. At least this is my experience leading sales and marketing teams trying to apply AI to their jobs. Now I've built a company around helping GTM teams do this (nativeGTM.com)
Love these extremes and it reads almost like a meme! So true! First cohort might say uuuuh maybe I should define the process first so I can „act“ 😄 (as if) or „gross“, me and structure? 🙈 About the second cohort, I made a „playable comment“ of them.. had to find another perspective than flagging the AI „bingo“ here on linkedin. Pretty simple: arcade games meet opinion and I’ll make sure to put the „15-minute task“ in the backlog! Will be fun. (no tracking: natvernier.com/play hope you enjoy!)
There's huge value in exploring a new tool Rachel Woods, even for 6 hours. The immediate win is automation and time saved. The long‑term benefit is learning, seeing something new, and figuring out other areas to apply your knowledge to. After all, we’ve sat through far longer and far more boring trainings in our careers. I like your idea of dedicating 10% of your time to consistently trying new tools. I’d just add that not everything will work on the first try, and that experimentation (and failure) is just as important as the successes. Ref. 1,000💡
The 10% rule is smart, but it assumes you can estimate the ROI upfront every time. With AI, most of us can't, bc we don't know what we don't know. I tried automating client onboarding last quarter. What I scoped as a 2-hour build turned into a week of figuring out what I didn't know. That's the real trap. Not the 6 hours, but not realizing you're still in discovery mode, not build mode. Honestly that might be the best argument for the 10% approach. You're not buying automation. You're buying learning, which ultimately makes you better at predicting the ROI next time.
The 10% rule is smart. But I'd push back on the 6-hour trap being about over-engineering. If it takes you 6 hours to automate a 15-minute task, you're probably using the wrong tools. That was true a year ago. It's not true now. The gap between "just do it manually" and "fully automated" has gotten smaller. The middle ground isn't about restraint, it's about picking better tools and learning basic product dev skills .
This is exactly it. The teams getting real results are running a constant loop: test, measure, fix, iterate, then report back. They add guardrails as they go and keep a written operating log so learning compounds instead of disappearing. Every staff member should be taught this operating rhythm now, not later. If a team is still waiting for the perfect moment, they are already behind. 🔁
I am not one to criticize going down the rabbit hole and exploring, spending 6 hours on something that may lead to a new thought, a transferable discovery or just simply learning. We are developing a workforce that depends on GPT to give them the answer if we stifle learning. In this day and age information and right answers are so easy to attain that true learning and the connections made through not getting it right the first time that lead to insight and experience are getting lost. The experience and knowledge gap is growing rapidly. So IMO, spend the six hours failing and learning on something inconsequential. That way, when you need to apply what you’ve learned to something more impactful or need to create a new solution you have that ability. Anyone can ask Claude how to do something. If you don’t add any more value than you know how to ask Claude, they don’t need you.
This is why Google’s new Studio function in Gmail was so eye opening to me. I used a template to create an action item in an email -> task created automation. It took less than two minutes. Maybe it only saves a minute here and there, 15 minutes a day …. But hey, that’s nearly an hour a week!
A question is: did the person who spent a day automating a brief task gain reusable skills to where, the next time, they can spend 15 minutes automating what would have taken a day, week, or month? And then another in 10 minutes? And another in 5? A billion years ago, I was helping businesses install their first PC and first accounting software. $20,000 for a small business to automate what they could easily do by hand? (A PC cost $5,000 back then) But ask for a report … and another report … and another report. Simply impossible by hand.