Welcome back to The Agent GTM Newsletter where we share actionable tips, free AI tools, and market insights every Friday. |
Everyone's talking about vibe coding.
Build your own app. Build your own agent. Build, build, build.
I've written about this a lot. It feels like magic that people who've never touched code can now ship working tools in an afternoon.
But there's a quieter move I think is under-utilized: plugging into agents other people have already built and tuned.
Today, there are thousands of agents, built by actual people who know their domain and spent hours refining that you just... use.
Part of what changed is the plumbing. MCPs, better integrations, agents that can actually plug into real tools and real data. You don't need to understand any of that to benefit from it. The stack got good. Someone else did the work.
So why wouldn't you build it yourself?
Here's a concrete example. I use a domain value estimator built by Dharmesh Shah. You plug in a domain and it tells you what it's worth, with reasoning. That's it.
Three reasons I use his and haven't tried to build my own:
- I don't actually know what makes a domain valuable. Length, keywords, TLD, comps — there's a whole framework I'd have to learn before I could even write a decent prompt. Dharmesh has already figured all of that out.
- I trust him. He's spent years talking and writing about domains, he's bought and sold plenty of them, he has actual skin in the game. When his agent spits out a number, I believe it in a way I wouldn't believe my own.
- And the thing I didn't realize until I'd used a bunch of these — he knows what to ask AI for. If I built my own, I'd ask vague questions and get vague answers. He's already figured out the right prompts, the right structure, the right output.
A few others I reach for regularly:
A Reddit scraper that summarizes what people are actually saying about a topic. I use it when I want a quick read on a topic — way faster than scrolling threads myself.
An earnings call analyzer where you can ask questions about a public company's earnings discussion. I don't want to read a 40-page transcript. I want to ask "what did they say about AI spend?" and get an answer.
One of my own that I believe is underutilized is the expert opinion researcher. Say I want to know what leading voices are saying about... sauna therapy or something. It'll go out, grab interviews from the key people, summarize their perspectives, and cite reference material. I use this a lot when I'm writing "thought leadership" content for clients.
None of these are things everyone needs to build themselves. Not for lack of ability — but because someone else has already done it better, and they put a ton of care into it.
We're all going to end up with a mix — some agents we build, some agents other people build. I just think most of us are over-indexed on the first half and sleeping on the second.
I've been thinking about this a lot lately. Working on something in the space. More soon.