internal ai tool
role
SDR
year
2026
timeline
3 days
team
solo
tools
n8n, python
OVERVIEW
prospect research eats the front half of an SDR's day — figuring out what actually makes a company buy. in three days i built an agent that does the first pass: pulls the signals, forms a hypothesis, hands back something sharper than a blank page. it won't always be right, but it doesn't have to be — it has to save the hour before the call.
CONTEXT
The problem is it often takes too long to research the GTM alpha of a company. What makes them buy? What makes them a customer?
If possible, why can’t we make an agent to help us to determine? (Obviously, not that it’ll always be right, but at least lessen the time)
FINAL ARTIFACTS
IMPACT
I don't have call outcome data yet as this was a build and not a deployment. But the design was intentional about what success would look like: if the SDR walks into a call with a sharper hypothesis than they'd have had otherwise.
REFLECTION
The hardest part of this project wasn't the code, but articulating what good SDR research judgment actually looks like, especially if one is specific enough that a model could follow it.
What counts as a disqualifier?
What makes an opening question genuine vs. leading?
When is a signal strong enough to act on?
(These come through experience as a SDR or asking tons of questions)