People love to ask me what my "stack" is, and I get why — it feels like if you just had the right tools, the business would run itself. It won't. Tools don't run a business; a system does. But the right tools remove enough friction that one person can do what used to take a small team. That's the whole game when you're lean.
So here's the honest version — not an affiliate dump, just what I actually open during a normal week and why. I've deliberately kept it small. Every tool I add is one more thing to maintain, pay for, and context-switch into. Fewer, sharper tools beat a bloated stack every time.
The rule I use before adding anything
Before a tool earns a spot, it has to clear one bar: does it remove real, repeated work — or does it just feel productive? A lot of software is the second thing. It gives you dashboards to stare at instead of hours back. I only keep tools that give me hours back.
A tool that saves you an hour once is a toy. A tool that saves you an hour every week is infrastructure.
1. Product sourcing & research
This is where I spend the least time now and it used to eat my week. I run an AI-assisted sourcing pass against a real supplier catalog — I want the tool to surface candidates against my criteria and, critically, to carry the boring fulfillment data (supplier, shipping speed, cost) so I can rule out the slow products automatically. A sourcing platform with a structured catalog like Doba is what makes that possible; the agent is only as good as the data it reads, and clean supplier + shipping data is exactly what lets it filter well instead of handing me a list I have to re-vet by hand.
2. Listings & copy
A general-purpose LLM is my writing partner here — titles, descriptions, bullet specs — but always as a first draft I edit, never as final output. The trick is feeding it your actual customer and your voice, not "write me a product description." Garbage brief, garbage listing. Good brief, and it gets me 80% of the way in a fraction of the time.
3. Customer support
An AI assistant handles tier-one support — order status, shipping times, returns, sizing. That's ~80% of tickets, and it's exactly the 80% that used to drain my motivation. I only touch the messages that genuinely need a human. This single layer is the difference between a store I dread and one I can run alongside everything else.
4. Ads & creative iteration
AI helps me generate ad variations and, more importantly, read the results faster than I can by hand — which angles are landing, which to kill. I still make the calls. The tool just lets me test more, faster, without drowning in spreadsheets.
5. The glue: a place to think
The least glamorous but most important part. Whatever you use — a doc, a notes app, a simple board — you need one place where the system lives: your criteria, your SOPs, your "what worked / what didn't." The tools execute; this is where the judgment compounds. Skip it and you'll relearn the same lessons every quarter.
What I intentionally don't use
No sprawling automation suite wiring twelve apps together. No tool I check daily just to feel busy. No "AI does everything" platform that promises to run the whole business — because the moment you stop reading what the AI produces, the quality craters and customers feel it.
Keep the stack small. Feed each tool a good brief. Read everything it gives you before it ships. That's not a secret hack — it's just the difference between AI amplifying a business that works and AI amplifying a mess. Your judgment is still the most valuable tool in the stack, and it's the one nobody can buy.
