TL;DR
Lead-Taxi is our attempt to make B2B outreach genuinely individual. Instead of putting 100 companies into one campaign, every company gets its own research, message and landing page. We’re testing it on our own MILTON sales first: roughly 100 target companies, with five qualified meetings as an interesting result and ten as a very good one.
Personal is more than a variable
Most outbound software is built around one basic idea: take a campaign and make it scale.
Upload a list, write a sequence, add variables, automate follow-ups. AI made this dramatically easier because now even the text can look individually written.
I never really liked that logic.
If I want to sell something to a specific company, I want to know why I am contacting that company. What is happening there? Who am I talking to? What could actually be relevant to them? If the answer is basically the same for everyone on a list, adding the company name to the first sentence does not make the outreach personal.
One company, one campaign
So Michael and I started building Lead-Taxi around the opposite idea: one company, one campaign.
For every target account, Lead-Taxi does the preparation individually. Research the company. Look for relevant signals. Prepare the approach. Write the message around that specific account. Create an individual landing page that continues the same conversation instead of sending everybody to a generic product page.
We have a working version now.
100 companies. A real test.
And instead of turning it straight into another SaaS product, we’re using it on ourselves.
MILTON is a good test for this because I know the market, I know who we want to talk to and I know what a useful meeting looks like. For the first real run, we picked roughly 100 target companies.
The success criteria are deliberately uncomplicated.
Five qualified meetings would make me interested. Ten would be very good.
If we send 100 genuinely individual approaches and almost nobody wants to talk to us, I do not want to explain that away with “we just need three more features.” Maybe the positioning is wrong. Maybe the research is not good enough. Maybe individual landing pages do not matter. Maybe the whole idea simply creates less value than we think.
All of those answers are useful before spending another year building the product.
Before the CRM
One thing also became clear while building it: Lead-Taxi should not become another CRM.
I do not need another database for contacts, opportunities and meeting notes. There are plenty of good systems for that, and we can connect to them.
The interesting part is what happens before somebody becomes an opportunity.
Who is actually worth approaching? What do we know about them? Is there a reason to contact them now? What should we say? What information would make the conversation useful for them?
AI can generate the email. That part is getting cheap.
Getting the premise right is harder.
For me, a simple test is: Would I still send this if I had to write it myself?
If the answer is no, an agent producing it faster does not solve much.
