I Built a Lead Generation Solution with GitHub Actions

As Head of Product, I'm responsible for our portfolio of products. We have several AI products and AI infrastructure. One of these includes a 30-day trial. When the user signs up, they get 30 days. This particular product is not SaaS. It's LLM keys and VectorDBs, so typical analytics as you would use in a SaaS to monitor adoption and conversion don't necessarily apply. That doesn't mean we can ignore metrics. Our sales, marketing, and product teams need to know whether people are using the product, what's working, and what leads to conversion. This is one of our newer products and we're working lean. Getting information out of where it does live and into the hands of our teams is what I set out to solve this week with AI tools.

On top of that, this product runs across six global regions. Each region has its own API tracking usage data. Meaning we have hundreds of trial users scattered across six separate systems. Doing this manually would be a nightmare.

What I built

A GitHub Actions pipeline that:

• Queries all six regional APIs weekly
• Filters out the noise (anonymous trials, test accounts, internal users)
• Aggregates spend per trial
• Exports to CSV
• Webhook sends to N8N
• N8N workflow makes its way to HubSpot

Zero infrastructure. No servers to maintain. Just a scheduled job that runs every week.

Teams see the final product, a workflow they can use in HubSpot. They assume it's sophisticated. It's not. It is effective. The teams can prioritize outreach based on who's using the product rather than guessing from the signup date.

That's the job. Notice the gap. AI empowers me to build the solution. Debug the weird errors. Ship it.