Today we’re introducing the world's first AI Workforce for Marketing Ops: a team of AI agents that converts 30% more inbound into demos/ signups.
Marketing teams spend millions on inbound, but 50%+ of leads never convert b/c they're lost to long forms, poor targeting, and slow follow-ups.
Here’s how the AI Workforce works →
🧵
vibe coded a CRM with Claude in a weekend. contact records, pipeline stages, deal tracking, the whole thing. felt like a genius.
lasted about 2 weeks before I realized I made a terrible mistake.
so here's what happened. we're a small team, 2 sales reps, and I convinced myself we didn't need HubSpot. why pay for something Claude can build in 48 hours right?
Claude built it. it worked. I was walking around the office like I just saved us $15K a year in software spend.
then week 2 hit.
a routing rule broke. nobody knew how to fix it except the engineer who built it. he was heads down on product. the lead just sat there unassigned for 3 days. nobody even noticed until the prospect emailed us asking why nobody followed up.
that was the first crack.
then we tried to hire a sales rep. first question in every interview: "what CRM do you use?" when I said "oh we built our own" the energy in the room just died. every single candidate knew HubSpot or Salesforce. nobody wanted to learn some homegrown tool that one engineer built in a weekend. I watched 3 great candidates lose interest in real time.
that was the second crack.
then I tried connecting Apollo. then Clay. then our email sequences. every single tool in our GTM stack has native HubSpot and Salesforce integrations. connecting them to our vibe coded CRM meant custom API work for every. single. one. more engineering time. more things breaking at 2am that nobody knows how to debug.
that was the third crack and I was done.
the thing nobody tells you about vibe coding your business tools is that building the V1 is the easy part. like genuinely easy. Claude will build you a beautiful CRM in a weekend. the hard part is everything after that:
who maintains it when something breaks at 11pm on a friday
who onboards new hires onto a tool with zero documentation
who builds and maintains every integration manually
who fixes the schema when your AI agent corrupts a field
the answer to all of those is "you" and eventually you get tired of being the answer to all of those.
established CRMs feel like overkill when you have 2 reps. but the migration you're avoiding now costs 3x more when you have 10 reps. ask me how I know.
two weeks in I got tired of explaining our homegrown CRM to sales candidates who just wanted to use HubSpot.
so we went back to HubSpot.
hubspot wins again 🤷♂️
Today we launched OpenClaw for Sales.
It combines @openmartai’s data, LinkedIn data, and top data sources to do the work for you.
Used by teams like Whatnot, DoorDash, Alibaba, and many others.
Try for free: https://t.co/WdJe81mD8j or DM me for an invite code.
I got into @ycombinator solo
After 7 rejections
Before the batch:
-$2.1M raised alone
-Sold 200K ARR
-6M+ views
I'm building the first database sandbox @ArdentAI
We let you infinitely clone any Postgres DB in <6s so coding agents can test code on a 1:1 of prod
Time to win
We've raised $6.5M to kill vector databases.
Every system today retrieves context the same way: vector search that stores everything as flat embeddings and returns whatever "feels" closest.
Similar, sure. Relevant? Almost never.
Embeddings can’t tell a Q3 renewal clause from a Q1 termination notice if the language is close enough.
A friend of mine asked his AI about a contract last week, and it returned a detailed, perfectly crafted answer pulled from a completely different client’s file.
Once you’re dealing with 10M+ documents, these mix-ups happen all the time.
VectorDB accuracy goes to shit.
We built @hydra_db for exactly this.
HydraDB builds an ontology-first context graph over your data, maps relationships between entities, understands the 'why' behind documents, and tracks how information evolves over time.
So when you ask about 'Apple,' it knows you mean the company you're serving as a customer. Not the fruit.
Even when a vector DB's similarity score says 0.94.
More below ⬇️
An engineer at Anthropic wrote a spec, pointed Claude at an Asana board, and went home. Claude broke the spec into tickets, spawned agents for each one, and they started building independently.
When the agent is confused it runs git-blame and messages the right engineers in Slack. By Monday the agents finished the plugin feature.
That's one example of how the best engineers are shipping software right now.
Developers will soon orchestrate 50 AI agents in parallel and the difference between a good engineer & a great one would come down to specs.
You can't write a spec that holds up at that scale without genuinely understanding what you're building at a deeper level.
The next-gen developer who understands the fundamentals, can architect well and orchestrate agent is going to be a 1000x developer!
@ShubhAgrawal26@Lovable@ShubhAgrawal26 let’s get on a call and i’ll show you claude code vs lovable; have some new features in surface also in wanna telll u about
hiring PSA for early-stage founders:
there are two types of employee candidates
1. ones who want to work at a proven rocketship
2. ones who want to help you build the rocketship
we've never gone wrong hiring the second type