AI & Operations
Here's the exact $70k/month machine I replaced with AI, and how.
Jake McKeon
April 2026 · 7 min read
At the peak of Coconut Bowls, I had 10+ people on the team. Today, I run a business on $2M+ with zero employees. I explain how I got here in the story of the two-year pause that rewired how I build.
Not because the team wasn't valuable. They were. But most roles were actually bundles of repeatable tasks. And AI is very, very good at tasks.
10+
People on the team at peak
$2M+
Current annual run rate
0
Employees today
The team I used to have
At our peak, the team looked like this: Wholesale and business development manager. Social media manager. Content coordinator. Community manager. Performance marketing manager. General manager. Two customer support staff. A social media assistant. And a personal assistant for me.
Most of them were hired for attitude over deep expertise. They were great people, aligned with the brand, and we built a culture I'm genuinely proud of. But here's what mattered in hindsight: they were executing systems I designed, not building their own.
That's exactly what AI replaces first.
What AI now does, role by role
Customer support
Fully replacedAI drafts every response, handles everything from order tracking to refunds. I review and send from drafts. Total time: 20-30 minutes a day for 10-20 enquiries, with very little editing needed.
Wholesale & business development
Mostly replacedAI handles outbound outreach and introductions. Wholesale customers can purchase without any hand-holding. Support routes through the same automation.
Social media & content
Rebuilt entirelyWas costing $5K-$10K/month on UGC creators. Now automations generate 10+ new product photos daily into lifestyle situations. World-class output. Essentially free.
Posting & scheduling
Fully replacedAutomated through a chatbot. I decide what goes out. I don't touch how it goes out.
Paid ads (Meta & Google)
Heavily augmented~70% of revenue comes from paid ads. AI writes copy, generates tests, builds campaigns, and publishes ads. I provide creative direction. Used to be a full-time role. Now about an hour of my day.
Email marketing
ReplacedAll flows, welcome series, abandoned cart, post-purchase, generated with AI. Weekly campaigns and optimisations executed with AI. Maybe 1-2 hours a week.
Personal assistant
ReplacedAI has access to my Google Drive, emails, and docs. Helps me find information, draft responses, and execute tasks.
Website development
ReplacedWas spending ~$2K/month on contractors. Now I use AI tools to build widgets, improve UX, and ship changes that directly impact conversion rate. Faster, better, no dependency on anyone else's schedule.
How much am I actually saving?
At the peak of Coconut Bowls, payroll was roughly $60,000-$70,000 per month. Call it $700,000-$800,000 a year.
$70K/mo
Peak payroll
$4K/mo
AI tools today
~95%
Reduction in operating cost
That's roughly a 95% reduction in operating cost. $55,000-$65,000 per month going back into the business. Capital that used to be locked into salaries is now fuel for growth. I can operate with thinner margins. I can spend more aggressively on ads. I can reinvest faster.
Previously, 10-15% of revenue was absorbed by salaries alone. That's gone.
What surprised me most
Two things.
First
What AI is already world-class at
The photos and videos it generates are genuinely competitive, not "good for AI", actually good. Pattern recognition, content generation, iteration speed, all of it has reached a level I didn't expect this soon. February 2026 was the real inflection point for me. That's when things went from interesting to genuinely useful.
Second
What it still can't do
It doesn't initiate. It doesn't decide what matters. It doesn't prompt itself. It still needs the human in the loop to identify what needs to happen and feed it the right information.
My operating system
Give AI better inputs than you think it needs
Actually read what it gives back. Most people skim, then say AI isn't good
Save what works so you can reuse it
The real trade-off
There is a catch. You become the system.
I am the operator, the decision-maker, and the bottleneck. If something puts me out of action for a stretch, the business doesn't collapse, but it slows down. That said, I can take two weeks off and the business runs fine with maybe an hour a day of oversight. No team morale to manage. No direction to set. Just maintenance.
For where I'm at in life, with young kids, that trade-off is worth it every time.
The part that really changed my thinking
With a team, scaling down is slow and expensive.
When Coconut Bowls peaked and started declining, margins compressed and competition increased, but I still had the weight of a large team. It took 12-18 months to properly scale back. It probably cost me a million dollars in delayed decisions, because you hope things will turn around, you delay the hard calls, and you carry the cost longer than you should.
Traditional businesses
Built for stability
12-18 months to scale back. Delayed decisions. Costs carried longer than they should be.
AI-native businesses
Built for speed and adaptability
Scale down instantly. Day by day. No lag. No emotional overhead. No redundancy costs.
That's not a cost saving. That's a completely different risk profile.
Where I'd tell founders to start
If you're going to replace one role first, start with customer support.
It was the first hire I ever made. It's now the easiest to automate. But more importantly, it keeps you close to the customer. Even now, reviewing support daily gives me product feedback, operational insights, and conversion signals that I'd never see if someone else was handling it.
I don't love doing it. But it's the highest signal, lowest cost role to transform. And the feedback you get from being that close to your customers makes every other part of the business stronger.
If you're on the employee side of this shift, I've written a separate piece for you.
The biggest risk in my old business was carrying too much team.
The biggest risk in my current business is just me. I'll take that trade every time.