Marketing
OpenClaw AI.
A customizable marketing AI framework that helps compare advertising messages, analyze email performance, and make sense of campaign results.
Built on real-world experience with AI-assisted marketing. The public project includes reusable agent instructions, Python analysis tools, and clear example outputs.
Which customer benefit attracted the most interest?
Imagine sending three emails to 1,000 people each. Each email communicates a different benefit: discovering something new, feeling confident about a purchase, or saving time. The tool compares how many people clicked each message.
What it means: In this example, the email promising discovery received the most clicks. It is a promising message to explore further, not proof that it will always work best.
Illustrative, fictional figures. No real customer or client data.
Practical marketing analysis, explained clearly.
The framework helps translate marketing questions and spreadsheet exports into repeatable analyses that a business owner or marketing team can understand.
Jobs-to-be-Done analysis
Compare different customer benefits—like convenience, confidence, or discovery—and see which message attracts more interest.
Explore analysis code ↗Email marketing analytics
Review which email campaigns generated clicks, orders, and recorded revenue using a simple CSV export.
View Python tool ↗Reusable agent instructions
Document how an assistant should approach research, marketing experiments, permissions, and analysis.
View agent instructions ↗From campaign numbers to a useful takeaway.
The Python tools read basic spreadsheet data and highlight how email campaigns performed.
- See how many recipients clicked each message.
- Compare recorded orders and revenue across campaigns.
- Identify promising patterns without confusing correlation with proof of causation.
Illustrative figures only. Actual performance depends on the data supplied.
Designed to be understood and adapted.
The public repository separates agent behavior, reusable skill instructions, and independently runnable Python tools.
IDENTITY.mdAgent role and expertiseSOUL.mdBehavior and communication principlesAGENTS.mdOperating rules and safetyTOOLS.mdPotential integrations and toolsHEARTBEAT.mdApproved recurring-task templateskills/Agent analysis instructionsjtbd/Message experiment analysistools/Email marketing analytics
Adaptable to a new application
This generalized framework builds on a previously used AI marketing implementation. New uses require appropriate configuration, authorized data access, integration setup, and testing before deployment.
Potential integrations include Shopify, Klaviyo, Google Sheets, advertising platforms, and public competitor research. These are not connected by default.
Read integration notes ↗Building on independently developed AI work.
The Jobs-to-be-Done approach in this project builds on JTBDGPT, an earlier initiative documented publicly in 2024.
JTBDGPT: A Jobs to Be Done AI Innovation Expert
The original article documents an earlier JTBD-focused AI initiative. The broader OpenClaw framework evolved later.
Read the original article ↗Explore the project on GitHub.
Review the agent instructions, Python scripts, synthetic datasets, and reusable skills that make up the framework.