- AI & Automation
- DevOps & Infrastructure
Admissions email AI agent
An AI agent that helps an admissions team answer applicant emails faster and more consistently.
- Client
- Tech training institution, East Africa
- My role
- Tech lead and automations developer
- Team
- Me and a contract developer
- Stack
- GitHub Actions nginx OpenAI
This project was delivered for a confidential client. Names, data and identifying screens have been removed or recreated.
The context
An admissions team handles enquiries from prospective learners across several programmes and intakes. Most arrive by email.
The problem
Many emails ask the same questions about programmes, intake dates and requirements. Answering each by hand takes time, and answers vary depending on who replies.
My role
I led the project technically and worked alongside a contract developer on the agent. I owned the deployment pipeline, the server setup and the project documentation.
What we built
An email agent powered by a large language model, paired with an admin application for the team. The admin app is served behind nginx and deployed through GitHub Actions, so every change goes through the same tested path to production. A master document index keeps the specifications, decisions and runbooks in one place.