teams that are solely relying on email for outbound are leaving so much on the table.
On LinkedIn we CONSISTENTLY see 8-12% reply rates across tens of thousands of campaigns.
and before you say "but you can't scale it", think again.
I built an AI agent that replies to my team's LinkedIn leads.
It reads every reply to our campaigns, drafts a response in our voice, and posts it to Slack for a one-tap approve. Approved → sent back to LinkedIn.
Here's how it works 👇
The problem: outreach works, but the inbox is chaos.
Hundreds of replies — "thanks for connecting," emojis, spam, and the occasional real buyer. The buyers get buried. Replying by hand = hours. Replying with a template = ignored.
So it doesn't blast everyone. It triages:
→ only people who replied to MY campaigns
→ skips spam, "not interested," and random inbound
→ flags the hot ones
Signal, not noise.
The part that surprised me: relevance.
It reads the full thread + the lead's profile and writes like someone who actually read it. One lead said "I'm not sure I ever used that tool" — instead of bulldozing the pitch, it adjusted.
It even matches each teammate's voice — my casual lowercase vs a colleague's warmer style. Same offer, different person. Leads can't tell it's assisted.
The stack:
• Aimfox (LinkedIn outreach) → webhook on new replies
• Claude (Opus) reads + drafts
• Slack card → ✅ approve / ✏️ edit / ❌ skip
• Approved → sent back to LinkedIn
A human is always the last step.
Result: We open Slack to ~50 ready-to-send drafts, approve the good ones in seconds, and book calls — zero copy-paste. Warm leads stop dying in the inbox.
Built the whole thing with Claude Code in 30 minutes. Happy to break the setup down further — reply and I'll share.
Are native integrations between tools becoming less relevant, now that everything is centered around AI?
I feel like founders don’t need to chase down all of these partnerships anymore.
All you need is to build a robust api and mcp
Fact: over half of our entire userbase at @aimfoxofficial is using our api and mcp
Are n8n and make finished?
I spent hours trying to buid a reporting system that ties in different sources and sends a slack message.
After like 8h i just gave up.
Then I did it with claud code in 25 minutes.
Last Wednesday I gave up on my new startup MoltSets entirely. I was devastated. Today, I think it's a $10m ARR biz because of ONE CHANGE we made to the offer.
Here's what happened:
First, a little background...
- 7 weeks I stopped working on RB2B
- To focus on MoltSets - unlimited contact data APIs
- We’ll be in closed beta next week - waitlist below
My original hypothesis for MoltSets is that a large number of Clay Agencies (maybe 200?) would pay $495/mo for MoltSets.
That alone isn’t an attractive business.
But I hoped how loud they were would pull in growth marketers and Founders at $199/mo, and individual reps who wanted to do sales in Claude at $19/mo.
Then I had 3 of the biggest Clay Agencies say the same thing last week:
“There’s probably 10-30 people who would pay $500/mo for that.”
FACK!
But then one of them suggested something WILD.
What if instead of trying to get HUNDREDS of people on $497/mo…
You tried to get THOUSANDS of people on a $97/mo plan that was unlimited and STILL had huge rate limits???
As I continued my conversations with GTM nerds, the enthusiasm for the $97/mo offer became palpable.
So much so, that a few people said words that are music to my ears:
“How can you run a profitable business with a $97/mo truly unlimited offer?”
When I explained it to them, it clicked:
1. We have incredibly cheap, massive awareness through LinkedIn
2. We are already using our database and data team (profitably) to serve three other business lines
3. MoltSets piggybacks on those two “moats” for another path to monetization of both of them, and
4. We will do this without humans, like RB2B
“Ah, I see.” (they said, nodding)
TAKEAWAY
Customer Discovery is more important than ever.
I’ve never sent a cold email or made a cold call, and I’m trying to design a product for people who are the best in the world at it.
Vibe-coding makes you think you should just throw a product out there and see what happens.
But unless you are LITERALLY eating your own dogfood and solving a PAINFUL PROBLEM that you actually have, there’s no way your vibe-coded slop will be disruptive.
Going $0-$1 quickly requires two things:
1. A disruptive product
2. Disruptive marketing
If you don’t have both, you’ll get stuck. I know because I’ve been there.
There’s only ONE WAY I know of to create both those things …
CUSTOMER DISCOVERY.
The market knows. You just have to ask.
P.S.
Join the MoltSets waitlist here: https://t.co/9CyzR4JIPS
Get Unlimited Contact Data APIs for Clay and Claude Code.
(We are trying something new w/ whatsapp, so if you signed up before, sign up again)
One of Aimfox users getting almost 300 replies in a single day with rented LinkedIn profiles. Just WOW.
Honestly, as a founder there's no better validation you can get.
I know we're building something huge.
and it's getting bigger every day.
That’s what we initially tried, but when you’re starting out and don’t have brand awareness it’s hard. There are just so many software out there and people just don’t care enough to try ever one. Regardless of whether it’s free or paid.
Best combo is having something like an experts program, were people use your software, pay you (a discounted rate), you give them promotional space through your website, blog, public templates, webinars, etc. and they represent the brand.
This is something that we started rolling out now. We get like 200-250 free trials started every day, 95% of those people are just doing this for them selves and would be intetested in learning about a DFY thing, we want to let our experts monetise the dfy potential.
Exactly - if you leave the note blank more get accepted, but that same drop will happen as soon as you send the first message.
Doesn’t change the end goal that much.
your linkedin connection request message doesn't matter as much as you think
tested 500 requests with message vs 500 blank
with message: 9% acceptance
blank: 21% acceptance
12% difference
what actually matters: mutual connections, profile strength, and whether you engaged with their content first
stop overthinking the 300 character note