SEAM Outreach
An AI outreach system that researches each prospect and writes messages worth replying to, at volume.
The problem
Generic AI-written outreach is worse than no outreach — recipients recognise it instantly and it burns the domain along with the prospect. But genuine personalisation means researching each company individually, which is exactly the work that does not scale. The interesting problem is not generating text; it is assembling enough real, specific context about a prospect that the generated message could not have been sent to anyone else.
What we built
- 01
An enrichment stage that gathers real signals about each prospect before any text is generated, so the model is working from facts rather than inventing plausible ones.
- 02
Message generation constrained by those signals, with the specific detail that justifies the outreach carried through into the copy.
- 03
Campaign and sequence management with per-prospect state, so follow-ups reflect what has already been said.
- 04
Response tracking feeding back into which framings actually earn replies.
- 05
Lead sourcing and cleaning pipelines feeding the top of the system.
Decisions worth noting
Enrichment before generation
The quality ceiling of an outreach system is set by context, not by the prompt. Most of the engineering sits in the enrichment stage, which is the opposite of where teams usually put it.
Operated, not just shipped
This runs live campaigns for real engagements, which means the feedback loop is reply rates rather than a satisfied stakeholder.
More work
VetBid
An AI business-development platform that reads federal solicitations and tells veteran-owned firms which ones they can actually win.
Next Door Notice
A geospatial monitoring platform that watches every planning application in London and tells people when one lands near them.
AwesomeGene
A fully automated video production pipeline: one brief in, a finished, published video out, with no human in the middle.
Tell us what is slowing you down
A short conversation is usually enough to tell whether this is a build, an automation, or something you should not do at all. We will tell you which.