Friends, family & referrals
coming soonWarm intros from people who already trust you. The best leads you'll ever get, and the first to run out. You only know so many people.
ex · Boardy
Manifesto
Every startup needs two things: a product, and customers.
For the last two years almost all the leverage went to the product side. Claude Code, Codex, Cursor. Building went from months to days. One person can now ship what used to take a whole team.
Getting customers didn't get that same leverage. If anything it got harder. Everyone can build now, so reaching the right people and actually standing out is the real bottleneck.
That's the side we're building. The way founders use Claude Code to build their product, they'll use Bavlio to get their customers.
Customer acquisition 101
Every way to get customers fits on one map: how warm the audience is, and whether you reach one person or many.
one to oneone to many
warmcold
Content
Where Bavlio livesBuild an audience once, reach it every day. It compounds, but slowly. It takes months before strangers start showing up.
- Text: posting on X & LinkedIn
- Short form: Reels and TikTok
ex · Stanley AI
Cold outreach
Where Bavlio livesFind the right person, learn what they care about, and write to them directly. It needs no audience and no budget, just effort. That effort is what Bavlio automates.
Run through the channels where people already are. Email and LinkedIn is where Bavlio started. WhatsApp, SMS, and socials like Instagram and Facebook are being added.
ex · Lemlist
Ads
coming soonPay to be seen by many. Works once the economics work. Most early products can't outbid the incumbents.
ex · StackAdapt
Bavlio started in cold outreach, and we're currently dogfooding content.
Ads and warm referrals are next — we're building into them later this year and early next. The goal is to cover the entire map.
The bet underneath it.
AI is not going to rebuild Google, or Stripe, or Supabase, or Vercel. The application layer keeps getting cheaper to copy. The infrastructure layer is what lasts.
Outreach is not as easy to copy as it looks. You can put together parts of the process in a weekend. You cannot put together the whole thing. Automating a single channel is doable. Running several of them at once is genuinely hard, and running them safely is harder still, because if you get it wrong the accounts get banned. You only have one LinkedIn, one Instagram, one sending domain.
So we built the parts that are hard. BaviMail is our own email infrastructure, and the sender reputation underneath it takes years. The sequencing engine on top of it is what runs a campaign across email, LinkedIn, and every channel we add next without putting those accounts at risk.
A single channel
One channel, automated. That part is not hard.
the application layer: cheap to copy
BaviMail + the sequencing engine
Our own email infrastructure, and the engine that runs multiple channels at once without getting the accounts banned.
the infrastructure layer: what lasts
Our own model, Bavi-1
A compound AI system that gets better with every campaign, tuned on real outcome data through an eval gate. Not a static prompt you set up once. It learns what actually gets replies, and it compounds.
Bavi-1 learns from every campaign that runs on Bavlio, not only from yours. That is the part you cannot replicate sending on your own.
- Campaign goes out
- Outcomes come back
- Eval gate scores them
- Bavi-1 learns
- then back around. Every cycle makes the next campaign better.
Where we are.
Still early. We're piloting and refining with a handful of founders right now.
We think every business will eventually run customer acquisition this way. We're building the engine that gets them there.