The simplest framework I’ve seen for getting more out of AI — from Dharmesh Shah, co-founder of HubSpot. 60% Repetition, 30% Iteration, 10% Experimentation. Here’s how to actually apply it, with examples for each.
The Source
Dharmesh Shah introduced the 60/30/10 rule at HubSpot’s INBOUND 2025 keynote and published the deep-dive on his own newsletter at simple.ai on September 4, 2025. It’s the cleanest mental model I’ve found for thinking about your relationship with AI.
The premise is simple: most people use AI in one of two failure modes. Either they reach for it only when it’s novel (no compounding) or they hammer the same 3 prompts forever (no growth). The 60/30/10 split fixes both.
60%
Repetition
The AI use cases you already know work reliably. Prompts and workflows you’ve tested, refined, and can count on every time. The bread-and-butter that saves you time every single day.
What this looks like for me
• Summarizing articles into 5-bullet briefs
• Drafting email responses I’ll edit before sending
• Brainstorming a list of angles for a content idea
• Pulling key insights from a meeting transcript
• Reformatting a doc, deck outline, or table
Why this is the biggest bucket: the value of repetition isn’t the novelty — it’s the compounding. If you save 20 minutes a day on these stable workflows, that’s 100 hours back per year. The boring stuff is where the leverage actually lives.
30%
Iteration
Taking your proven use cases and systematically making them better. Same tasks, evolving prompts. Not changing what you’re asking AI to do — refining HOW you ask to get measurably better results.
What this looks like for me
• Adding a quality rubric to my caption-drafting prompt
• Testing 3 different opening lines for my morning brief
• Adding a “before answering, find one weakness in this” instruction to my decision-prep prompt
• Replacing “summarize this” with the 4.7-compatible “give me a full breakdown, not a summary”
• Building a custom Project so the prompt has standing context every time
Why this is critical: the people who win at AI aren’t the ones with the most use cases — they’re the ones with the most-refined use cases. Iteration is where the quality gap between “decent AI user” and “great AI user” gets built.
10%
Experimentation
Using AI for completely new applications you’re not sure will work. Dharmesh calls them your “what if” moments — the boundary-pushing tests. Most will flop. That’s the point.
What this looks like for me
• “What if I had Claude write a script for a podcast episode I haven’t recorded yet?”
• “What if Claude analyzed my last 12 months of decisions and told me what patterns I keep falling into?”
• “What if I built a Skill that interviews customers via DM?”
• “What if I had Claude design the brand identity for an unannounced product?”
Why this is non-negotiable: the experiments that work become next quarter’s Repetition layer. The experiments that don’t teach you exactly where AI’s limits are right now — which is information your competitors don’t have. 10% is small but compounds.
The framework only works if you actually check your own ratio. Most people assume they’re doing all three. Almost nobody actually is. Here’s a quick audit you can run in 10 minutes.
01
List your last 20 AI sessions
Open your ChatGPT or Claude history. Skim the last 20 things you used AI for this week.
02
Tag each one
R = Repetition (an existing prompt/workflow). I = Iteration (a refinement of an existing one). E = Experimentation (a brand-new attempt). Be honest.
03
Compare against 60/30/10
If you’re heavy on Repetition with zero Iteration, your AI usage isn’t compounding — you’re stuck. If you’re heavy on Experimentation with no Repetition, you’re chasing novelty and not banking the workflow time. Adjust accordingly next week.
04
Schedule the missing tier
If you need more Iteration, block 30 minutes weekly to refine ONE existing prompt. If you need more Experimentation, block 30 minutes weekly to test ONE wild new idea. Both compound.
The Reframe
Most people think “getting better at AI” means learning new tricks. The 60/30/10 split says it’s actually about structuring your time. Same hours per week, just allocated across the three tiers. Two months in, the quality of your AI work won’t look like the same person did it.
The Real Win
Repetition gets you back time. Iteration gets you quality. Experimentation gets you the edge. Run the split for a few weeks and your relationship with AI changes from “a tool I reach for sometimes” to “a system that compounds every week.”
The Only AI Masterclass You Need
If this guide helped, but you’re looking to go deeper, I got you!! My 30-Day Challenge takes you from saving AI tips you never use to actually building with AI, step-by-step.
I show you exactly how I automated two e-commerce brands, my social media, and most of my personal life, then hand you the agents, workflows & systems to do the same. I’m teaching you every single thing I know with one lesson and one build a day.
Join the AI Masterclass →© 2026 Mariah Brunner. All rights reserved.