Case study
Three demand engines, one pipeline, one launch.
A complete go-to-market system for an enterprise AI platform: outbound, inbound and partnership engines feeding one measured pipeline, aimed at a single launch moment.
Designed and operated by one person, in production at the AI startup where I run marketing. By Amber Lan, AI marketing engineer. Last updated 20 August 2026.
Back to selected workThe problem.
An enterprise AI platform two quarters from a hard launch window, with no marketing team.
No position, no engines
Positioning undecided, no message architecture, nothing running. Enterprise buyers research through analysts, peers, search and AI assistants, so a single channel was never going to hold.
One operator, no agency
Every system had to run solo: automated where machines are better, human only where judgment matters. An enterprise audience punishes overpromising, so every public claim needed a source or it did not ship.
What was built.
Three engines converge on one CRM pipeline, so every channel is measured against the same definition of a qualified conversation.
Outbound
Signal-triggered research and personalised outreach inside safe sending limits. Automation stops the moment a buyer replies.
Inbound
Search and answer-engine visibility, plus a content engine producing platform-native artefacts from single source documents.
Partnerships
A lighthouse program that gives early customers a co-marketing reason to go public.
Underneath: a claims register holding every public claim with its evidence and status, a message architecture mapping claims to buyer roles, and an operating timeline sequencing all of it toward one launch moment.
Key decisions.
Four calls that made the system operable by one person.
One launch moment, not a drip
Enterprise attention compounds when the category claim, the proof and the availability land together.
The claims register before the copy
Writing became fast because what could be said was already settled.
Engines that keep running
Systems over campaigns. Nothing in the plan stops working when a sprint ends.
Human takeover as a hard rule
The moment a person replies, automation stands down. Trust is easier to keep than to win back.
The result.
The plan lives as a navigable deck the leadership team uses, and the engines it specifies are in production.
In production
Positioning, message architecture, channel design and the operating timeline, run by the person who designed them, feeding one pipeline ahead of launch. Real numbers publish here once the launch window closes, not before.
Evidence
A sanitised case: the company's identity and internal numbers stay private, and the full walkthrough is available on a call. Public: the clean-room content engine behind the inbound side, and the role where the system runs today.
What I learned
Sequencing beats volume. The same assets, landed in order against one launch moment, do more than twice the assets scattered across a quarter. A claims register is the highest-leverage marketing document nobody writes.
Want a system like this pointed at your business?
Tell me what you do and where the pipeline stalls. You get a written read within the first week.