Workflow 02 / Outbound
Outbound Personalization Workflow
Researches an account, identifies relevant signals and creates structured personalization inputs a sequence can actually use.
The Outbound Personalization Workflow is a seven-step research process that produces structured personalization inputs for one target account: its business context, a named change signal and its source, the function that owns the problem, one checkable piece of evidence and a stated reason to contact now. It takes roughly 5–10 minutes per account, and accounts without a usable signal are skipped.
Why It Exists
Personalization usually means inserting a company name and a recent post. That is decoration. This workflow produces a specific, defensible reason the message is landing in that inbox this week.
Tools
Inputs
- Target account list
- ICP problem statement
- Offer or point of view
The Workflow
Run it in order.
- 01
Establish the account context
Business model, who they sell to, and how they make money. Relevance starts here.
- 02
Look for a change signal
Funding, hiring, leadership change, product launch, market entry, regulatory shift. Change creates the reason to act.
- 03
Locate the function that owns the problem
Identify the role most likely accountable for the outcome you affect.
- 04
Find the specific evidence
One concrete, checkable observation — a job posting, a stated priority, a visible gap.
- 05
Connect evidence to consequence
State what that observation likely means for them commercially. This is the part that earns a reply.
- 06
Produce structured fields
Output discrete variables the sequence can insert, rather than a paragraph a human must rewrite.
- 07
Quality gate
If the research produced nothing specific, the account is skipped rather than sent a generic message.
Outputs
What it produces.
- Structured personalization variables per account
- Named change signal and its source
- A stated reason-to-contact-now
- A skip list of accounts with no usable signal
Run It With Your AI
Install this workflow as an agent skill.
The same steps, packaged so your AI agent can run them with you: it asks for the inputs above, works through each step and produces the outputs in one document.
- Claude (web or desktop): download the .zip, then open Settings → Capabilities → Skills and upload it.
- Claude Code: unzip it into
~/.claude/skills/(or a project's.claude/skills/). - ChatGPT Codex: unzip it into
~/.agents/skills/(or a repository's.agents/skills/). - Cursor, GitHub Copilot, Gemini CLI and others: add the folder to that tool's skills directory (see agentskills.io (opens in a new tab)).
- Then just ask. For example: “Find a reason to contact these 10 accounts this month”. The agent loads the skill when the task matches.
Related Service
Automated Outbound
Build targeted outbound systems that identify the right accounts, reach the right people and create qualified conversations.
See the serviceRun It For You
This workflow is part of how we deliver Automated Outbound.
Use it yourself, or have us operate it as a system.
Make It Yours
Want this built around your stack?
This version is general on purpose. Ours get adapted to your CRM, data, ICP and team, and connected to the other workflows your GTM needs.
Full enquiry form