Sales Dev Agent

Earn the right to a conversation.

1Principles

ValuesBeliefsControls

Qualify before contact, not after.

Problem	Question	Decision
Volume, not quality	Can AI pre-qualify before handover?	Test with HRV via Sneakers Media

Agent scores prospect cards while human reviews outreach draft

2Performance

ValuesBeliefsControls

Five calls, fifteen percent, twenty contacts.

Problem	Question	Decision
No feedback loop	What's converting?	Wire channel dashboard

Dashboard showing reply rates, pipeline scores, and kill date countdown

3Platform

ValuesBeliefsControls

80% composed, 20% new code.

Problem	Question	Decision
9 algorithms, 0 demand	What connects with minimum new code?	Sprint -1: email + LinkedIn

Architecture diagram: 9 existing algorithms wired into one outreach pipeline

4Protocols

ValuesBeliefsControls

Research, score, compose, sequence, measure.

RESEARCH → SCORE → COMPOSE → SEQUENCE → MEASURE
    ↑                                        |
    └──── Feedback improves next cycle ──────┘

Score	Route	Action
Hot (>80%)	Direct	Personalized email + LinkedIn
Warm (50-80%)	Nurture	3-touch sequence
Cold (<50%)	Content	No direct outreach

Problem	Question	Decision
Same message to everyone	Route by fit score?	Adaptive explore-exploit

WorkChart: 5-stage loop with feedback arrow from measure back to research

5Players

ValuesBeliefsControls

Two ICPs, one WorkChart, two revenue paths.

Context	ICP	Job
Internal	Construction/solar EPC	Discovery calls for Stackmates
External	NZ agencies via Sneakers	Pre-qualified leads for HRV + clients

They Say	They Mean
"Lead quality is always poor"	"We get blamed after handover"
"We need to test it first"	"Show proof, not promises"

Problem	Question	Decision
Agencies buy leads, clients complain	Sell confidence, not contacts?	Pilot HRV

Split screen: internal dogfood pipeline left, Sneakers Media client pipeline right

1 / 5

Five Questions