Sales agent
It researches the account, scores fit, prepares context, drafts the next note, and leaves your team with a clean review step. No mystery and no auto-send.
Summary
Runs
Research, qualification, enrichment, pipeline review, draft follow-up, and win-loss analysis
Reads
ICP, offers, brand voice, pipeline files, CRM exports, meeting notes, and public sources
Writes
Profiles, scored verdicts, revival lists, pipeline notes, draft emails, and sales reports
Guardrail
No outreach is sent and no live CRM record is changed without a person
Where it helps
The first agent should have a clear job, known inputs, and a review path the team can keep using.
What it can run
These capabilities come from the public Conversion Skills patterns. The names under each card are the documented skills the build can draw from.
The agent can build a sourced profile from public information, then enrich an existing account or prospect file with company context, role context, recent signals, sources, and confidence notes.
It can compare a lead against the ICP, mark each criterion as met, missing, or unknown, produce a fit score, and recommend the next human action instead of guessing.
It can mine CRM and pipeline exports for stalled deals, dormant accounts, upsell candidates, and cold prospects, then rank the revival actions.
It can refresh deal stages, next steps, values, and stalled-deal signals from CRM exports, call notes, or user input, while leaving live CRM changes for human review.
It can write a one-to-one email from a real signal or a multi-step email sequence from offers, ICP, and voice. The output is a draft only.
It can analyze win-loss patterns, stage conversion, cycle time, rep performance, and sales coaching points from the pipeline and CRM export.
How it works
Step 01
The agent reads the ICP, offers, voice, pipeline, and any approved source files first, so the work starts from your business instead of a blank prompt.
Step 02
It researches, scores, enriches, drafts, or analyzes the specific sales job, then attaches sources and confidence where the repo requires them.
Step 03
The output names the next human action: review, send, book, ask for missing info, update the CRM, or drop the pursuit.
Human gates
The agent prepares work and keeps the trail clean. People still approve the moves that can affect buyers, customers, money, permissions, or public claims.
Outputs
The useful output is not a magic answer. It is a reviewable artifact, a clearer next move, and a source trail for anything factual.
Sourced prospect and account profiles
Lead qualification verdicts with reasons
Enriched pipeline records with confidence notes
Dormant account and stalled-deal revival lists
Pipeline health and win-loss reports
Draft emails or sequences a person reviews and sends
Related agents
Most useful AI systems start with one agent, then connect the next repeated job once the first one earns trust.
Public standard
Conversion Skills is the public reference for these patterns: plain files, documented skills, proof-aware outputs, and review gates. The agent pages show how that public method turns into useful business systems.