Xandrov StudioAtlas of shipped systems
← The territory/ II. Applied & Vertical AI/ № 07 — OutboundIQ
Dossier № 07 / 19 — Applied & Vertical AI — Applied AI · Go-to-market

OutboundIQ

Personalization from research, not merge fields.

File
№ 07 / 19
Territory
Applied & Vertical AI
Classification
Applied AI · Go-to-market
Status
Working generator
§ 01 Context

Cold outreach succeeds or fails on relevance. A lead list and an offer are the starting materials — but the gap between them and a compelling email is vast. OutboundIQ sits within the lab's lead-generation family of systems, filling that gap. It automates the research step that separates personalized outreach from templated spam.

§ 02 The problem

Template merge-fields (inserting a prospect's name or company into a boilerplate) read as spam. Hand-writing every email doesn't scale. The practical middle is a system that actually researches each lead — understanding their business, recent moves, fit to the offer — and then composes fresh copy under constraints. That middle path had no off-the-shelf solution.

§ 03 The approach

OutboundIQ is a Node + Express service wrapping Gemini. The interface is simple: a lead and an offer go in. The system runs a two-stage pass. First, it researches the prospect — gathering context from public sources, inferring needs and angles. Second, it composes an email under structural constraints: tone, angle, and messaging priority are set per campaign as configurable dials. The output is a fresh email, not a template with fields filled in.

Tone and angle are the levers. A campaign targeting finance teams might emphasize operational efficiency and risk reduction. The same offer to engineering teams emphasizes technical depth or implementation speed. The system doesn't swap adjectives — it rewrites the argument. Structure constraints (opening hook, body, call-to-action) ensure emails land in a defensible format.

Personalization is earned by research, not faked by merge fields.

§ 04 Decisions that mattered
D1
Personalization through research
Merge-field personalization is fake and reads as such. True personalization requires understanding the prospect — their business, their context, their likely objections. The system researches before composing.
D2
Prospect data never published
The system and its architecture are shown. Prospect lists, email outputs, and client identities are withheld. This preserves confidentiality and forces the focus onto the mechanism.
D3
Dials over templates
Rather than selecting from email templates, campaigns define tone, angle, and structure as parameters. Gemini composes a new email for each prospect within those constraints.
§ 05 Where it stands

OutboundIQ is a working generator. It exists as part of the lab's lead-generation family — a sibling to the 2-layer lead crawler. It solves the outreach side of the pipeline: taking a lead and an offer, researching the fit, and producing a personalized email at scale. It demonstrates that research-backed composition, not template substitution, is the foundation of relevant outreach.

Node + ExpressGemini
§ 06 — Sanitization record

Prospect lists, actual email outputs, and client project details have been removed. These materials would compromise confidentiality and shift focus from the system's architecture to the data it processes. The dossier shows how the system works; the leads and their attributes remain private.

Have a system like this worth building?

Fixed scope, quoted after one working session, first output inside two weeks. Read by a person within one working day.