Let’s talk AI.
AI doesn’t fail quietly. It fails expensively.
I keep seeing companies deploy tools before they decide what work should disappear, what work should speed up, and what still needs a human owner. The key to using AI effectively is finding that balance.
Here’s where I think you should and shouldn’t use AI, and what I’d replace first.
Let’s get into it.
Where to Use AI Right Now
Sales Development
Outbound: You can’t replace it. You can only augment it. AI works as a researcher, list builder, signal detector, and drafting copilot. If you try to fully replace humans here, you don’t get leverage. You get spam. Fast.
Inbound: You can replace most of it. AI can respond via email, phone, SMS, research the account, qualify intent, and book meetings. If it’s trained properly, it’s faster and more consistent than most teams.
Chat: You can fully replace SDRs here. Not button bots. Not workflows. Full conversation agents.
Sales
Transactional: You can replace a large portion. Simple pricing, clear use cases, low risk decisions. AI closes these already.
Relational: You can’t replace this. You can augment it with prep, insights, deal coaching, and follow-up. But trust, nuance, and pressure still need a human.
Marketing
Execution: You can replace most of it. Content drafts, repurposing, ops, reporting, visuals, landing pages. AI is already better at the volume game.
Strategy: You can’t replace it. Positioning, taste, judgment, and tradeoffs still matter. Most teams confuse activity with strategy. AI exposes that.
Customer Support
Tier 1: You can replace this. FAQs, password resets, order status, basic troubleshooting. Pure volume.
Tier 2: You can replace a lot of this. Known issues, repeatable fixes, guided workflows. AI handles this if it’s trained and owned.
Tier 3: You can’t replace this. Edge cases, bugs, emotional customers, high-risk situations. Humans should only see the hardest problems, not the volume.
Customer Success
Transactional: You can replace this. Anything that exists to move volume, push reminders, or prove activity instead of outcomes is AI work now.
Relational: You can’t replace this. Real problems. Real risk. Real money on the line. But if you’re staffed like every customer needs this, you’re bloated.
HR
Execution: Ops, screening, scheduling. You can replace most of it.
Strategy: Hiring judgment, culture, leadership calls. You can’t replace this. If you try, you’ll poison the org.
Engineering
Throughput: You can’t replace engineers, but AI multiplies their output. Faster cycles. Faster prototypes. Faster mistakes too.
Ownership: You can’t replace this. Design decisions, system integrity, and long-term bets still require humans. This is why headcount doesn’t drop. Pressure rises.
Bottom Line
AI replaces motion, not responsibility.
If your org confuses activity with ownership, this transition will hurt.
Where would I start?
If I were rebuilding a team from scratch, I’d probably start with Engineering, then move immediately to Customer Support.
Engineering comes first because it sets the ceiling. If your engineers aren’t AI-native, everything downstream is slower, sloppier, and more expensive than it needs to be.
Once engineering is equipped, I’d go straight to Customer Support. Why?
Support is where volume meets reality. Same questions. Same issues. Same friction. Over and over again.
This order matters:
Engineering first, to increase the rate of change
Support second, to eliminate noise and reclaim capacity
Most companies sprinkle AI on the edges, run pilots, and argue about “strategy” while the backlog grows.
More on this and other topics here.
That’s it for today. Connect with me on Linkedin if you actually want to understand what an Autonomous Organization looks like in the real world.
