How a Hong Kong government carpark automated 90% of public enquiries

A Genny-powered WhatsApp agent, backed by a custom RAG knowledge base, now answers thousands of repetitive public questions per week — freeing operations staff to handle real exceptions.

Before. Every WhatsApp message was triaged by hand. Repetitive questions dominated the queue, there was no after-hours coverage so the backlog arrived with the morning shift, and answers varied between operators.

What we built. Genny, our tool-calling WhatsApp agent, deployed on top of a heavy RAG knowledge base built from the carpark's own documentation — so answers are grounded in the operator's real rules rather than a model's guess, and every one carries source attribution for audit.

Outcome. 90% of public enquiries answered without a human, staff workload down 70%, and operators now see only escalations and genuine exceptions. Related: the WhatsApp agent and AI for the public sector.

All case studies · Talk to our team

Frequently Asked Questions

What is smart parking technology?

Smart parking technology uses sensors, real-time data, and software (often AI-driven) to monitor, allocate, and manage parking spaces automatically instead of relying on manual staff processes. It spans two layers: physical infrastructure (occupancy sensors, dynamic signage, automated payment) and communication/software layers such as enquiry handling. Genium's Hong Kong government carpark project sits in the second category — a WhatsApp agent backed by a RAG knowledge base that automates public enquiries without touching the physical parking hardware.

What is AI-driven parking management?

AI-driven parking management applies machine learning and automation to operational tasks like enquiry handling, occupancy prediction, and enforcement triage, cutting down manual staff workload. In Genium's HK government carpark case, an AI WhatsApp agent (Genny) paired with a custom RAG knowledge base automated 90% of public enquiries and reduced operations staff workload by 70%, leaving human staff to handle only escalations and genuine exceptions.

Can AI-powered parking management work with existing parking infrastructure?

Yes — AI-powered parking management can typically be layered onto existing parking infrastructure without replacing hardware, as long as it can access the operator's own rules and documentation. Genium's carpark deployment integrates as a WhatsApp channel backed by a RAG knowledge base built from the operator's existing documentation, so answers are grounded in real rules rather than a model's guess, with no changes required to the physical carpark system.

Does AI-driven parking management replace human enforcement teams and jobs?

AI-driven parking management does not eliminate human enforcement teams; it removes repetitive, low-value work so staff can focus on exceptions and decisions that require judgement. In Genium's HK government carpark case, automating 90% of routine enquiries reduced staff workload by 70%, but every escalation and genuine exception still goes to a human operator — the team's role shifted rather than disappeared.

How does AI-driven parking management improve EV charging availability?

AI-driven parking management can improve EV charging availability by using real-time occupancy and enquiry data to flag free chargers and answer driver questions about charger status without staff intervention. Whether this actually moves the needle depends on how directly the charging network's live status feeds into the underlying knowledge base — Genium's deployments in this space focus on that communication layer (RAG-backed WhatsApp agents) rather than the charging hardware itself.

Is AI-powered parking management only for large or mixed-use properties?

No, AI-powered parking management is not limited to large or mixed-use properties — the underlying technology scales down to a single site just as readily as a sprawling mobility hub. The determining factor is enquiry volume and rule complexity rather than site size; Genium's own case study is a single government-run carpark, not a mixed-use development.