Learn and master Clay with Clay university
Featured in Clay University
Use Case: Email Copywriting
In the last lesson, we covered two enrichment Functions: firmographic enrichment and company hierarchy.
This Function writes personalized outbound emails. You pass in an email address and get back three ready-to-send message variants.
You'll see how to use Functions not just for enrichment, but for taking action on that data.
📥 Step One: The Input
The Function takes one input: a contact's email address. Everything else gets looked up.
🔍 Step Two: Data Enrichment
The Function pulls intelligence about the contact from two directions.
First, a Get Contact step looks the person up in Salesforce to get their basic contact record and Account ID, which connects the person to their company.
Then two Snowflake queries run.
Contact Intelligence pulls first-party person-level data:
- Product usage patterns like workspace opens, table activity, and enrichment usage
- Conversation history from recorded calls
- Job change information and employment timeline
- Previous Clay experience
- Engagement scores and contact stage
Account Intelligence pulls first-party company-level data:
- Firmographics like industry, size, revenue, business model
- GTM motion and company segment
- Account-wide conversation data across all contacts at the company
- Pain points mentioned in calls
- Features discussed and feedback shared
- Competitor mentions
- Multi-threading opportunities
From one email address, the Function now has product usage data, CRM data, call transcripts, and competitive intelligence.
🧠 Step Three: Analysis
The Function takes all that data and makes sense of it.
A Persona Analysis step uses Claude to determine the prospect's persona: Sales, Marketing, Operations, RevOps, Growth, Executive, and so on. It identifies their seniority level, pain points, and the best messaging approach.
A Context Summary step creates a four to six sentence brief for the SDR. This prioritizes the strongest personalization angle, references conversation context, surfaces likely objections, and flags multi-threading opportunities.
✍️ Step Four: Message Generation
The Function generates three message variants.
An Insight-Led Message. Three to four sentences. Opens with a specific, personalized insight about the prospect, ties it to a relevant Clay capability, and ends with a soft conversational CTA.
A Value Nugget Message. Leads with a concrete, tactical insight and focuses on immediate value delivery.
A High-Signal Message. Two to three sentences. Leads with the highest-value qualifying signal: a job change, previous Clay experience, a competitor mention, a pain point from a recorded call, or recent product engagement. If none exist, it returns "NO HIGH SIGNAL - USE STANDARD MESSAGES."
📤 Step Five: Output
The "Send data back" action writes all three messages back to the original table as a JSON object. The SDR picks the variant that fits the situation, or uses an agent to make that choice and route the message to a Slack channel or directly into a sequencer.
One email address in, three personalized messages out.
🔮 What's Next
We've gone from enrichment Functions that pull data in, to a copywriting Function that writes targeted emails using product usage data and conversation history at scale.
In the next lesson, we'll cover signal detection and lead qualification.
