Project — AI Product Design & Engineering
An AI-powered receptionist that handles customer conversations, captures leads, and helps businesses turn messages into appointments.
Portfolio demo — a fictional dental clinic, a rule-based conversation engine, and simulated business actions. Nothing here books a real appointment or reaches a real person.
Capabilities
Answers common customer questions.
Captures customer intent and lead information.
Guides customers toward appointment requests.
Keeps conversations moving.
Knows when a human should take over.
Live Demo
This is a real, working conversation engine — not a scripted animation. Ask about hours, services, or pricing, try booking an appointment, or ask to reschedule or cancel one.
Simulated conversation — no real messages, bookings, or staff are involved.
Illustrative Customer Economics
Model the potential impact of recovering missed bookings, improving customer retention, and reactivating inactive customers.
Planning assumptions only — not actual customer results.
Recover
Potential annual revenue from missed bookings
$36,000
Retain
Potential annual revenue protected/generated through improved retention
$40,000
Reactivate
Potential annual revenue from bringing inactive customers back
$16,000
Recover
Missed bookings/month × Recovery rate × Avg. revenue per new customer × 12
Retain
Active customers × (Expected retention − Current retention) × Avg. annual revenue per retained customer
Reactivate
Inactive customers × Reactivation rate × Revenue per reactivated customer
Net impact & ROI
Total impact − ($300 × 12) = Net impact. Net impact ÷ $3,600 = ROI multiple.
These are illustrative planning assumptions, not actual customer results. Actual outcomes depend on message volume, booking conversion, customer value, retention behavior, follow-up effectiveness, and the business's existing processes. Recover, Retain, and Reactivate are modeled as separate customer populations (new bookings, existing active customers, and lapsed customers) — if your real numbers overlap between categories, adjust the inputs so the same customer isn't counted twice.