75% faster email replies
A retrieval-augmented email assistant that writes in the professor's own voice, turning hours of correspondence into minutes.
- Client
- Dr Laura Vella · University of Melbourne
- Built with
-
- RAG
- Pinecone
- Email automation
Academic inboxes never sleep. Between students, collaborators, and committees, Dr Laura Vella was spending hours each week on email, and every reply had to sound like her.
The challenge
Generic AI drafts miss the mark. They don’t know the history with a particular contact, they flatten tone, and they need so much editing that they rarely save any time at all.
What we built
A retrieval-augmented assistant grounded in Dr Vella’s own correspondence:
- Voice modelling: the system studies her past writing to match phrasing, structure, and tone.
- Context retrieval: a Pinecone vector database surfaces the relevant thread history and reference material for each message.
- Tone adaptation: replies shift automatically depending on whether she’s writing to a student, a peer, or an external body.
The outcome
Replies that used to take hours now take minutes, and the drafts genuinely sound like her, not a template.
“The email RAG system has been a lifesaver! It’s like having a personal assistant who knows exactly how I want to communicate.”
- Dr Laura Vella, Associate Professor, University of Melbourne
Most of these start the same way: a process that works, done by hand, too many times a week.