In development
AURA AI Receptionist
The problem. Service businesses — clinics, dental practices, salons, contractors, small firms — lose revenue to enquiries arriving when nobody can answer.
The approach. A three-channel agent covering voice, WhatsApp and website chat, sharing one conversation model so a customer is not restarted when they switch channel.
The governance. Escalation is designed in rather than bolted on: anything sensitive, contested or financial routes to an authorised person instead of being resolved by the model.
In development
AURA revenue recovery MVP
The problem. Businesses with a gap between an enquiry and a confirmed appointment lose a predictable share of that pipeline to silence.
The approach. A direct-booking and structured follow-up workflow that closes the gap without pressuring the customer.
The governance. The system does not discuss balances, disputes or payment terms with customers. Those conversations escalate to authorised staff by design — a deliberate constraint rather than a limitation.
Academic project
AI receptionist for real estate
The brief. A capstone project examining whether an AI receptionist is appropriate for a real-estate practice, and where its boundaries should sit.
My role. Requirements gathering, workflow mapping, solution design and evaluation of failure modes.
What it taught. The design questions that mattered most were about handover — when a conversation should stop being automated — rather than about conversational quality.
Consulting instrument
AI readiness assessment for SMBs
The problem. Small businesses are told to adopt AI without any means of judging whether they are ready or where to begin.
The approach. A structured assessment covering data, process ownership, staff capacity, risk tolerance and governance — producing a prioritised recommendation rather than a score.