Implementing an AI Chatbot for a New Zealand Medical Regulator
- Jun 3
- 2 min read
A New Zealand–based medical regulatory authority serves a large practitioner community and the general public. Its team receives a high volume of repetitive enquiries relating to Continuing Professional Development (CPD), professional standards, Annual Practising Certificates (APCs), and complaints processes.
To address this demand without increasing staff workload, the organisation engaged Zephyr Consulting to design and implement an AI-powered chatbot for its public website. The chatbot was deliberately scoped as a conservative, low‑risk first release version that focused on information discovery and enquiry reduction.
Challenge
The organisation faced several interconnected challenges:
High volumes of repetitive enquiries sent to a shared reception inbox
Seasonal enquiry spikes during annual renewal periods, particularly February–March
Users struggling to locate the information
Staff time was being diverted from higher‑value work to handle routine questions
A need to maintain accuracy, professionalism, and regulatory credibility
As a statutory body, the organisation required strict controls around risk, tone, escalation, and content sourcing, so the information had to be from the client's website and not reference external information.
Objectives
The primary goals were:
Reduce repetitive, low‑value enquiries reaching staff
Help practitioners and the public quickly find accurate, authoritative information
Improve overall website user experience through instant, consistent responses
Ensure all chatbot responses were based exclusively on approved website content
Introduce AI capability in a controlled, auditable, and regulator‑appropriate manner
Solution
Zephyr Consulting delivered a first‑line AI chatbot embedded on the organisation’s public website.
Key Characteristics
Information discovery focused: The chatbot directs users to existing web pages rather than providing new interpretations or advice
No internal system access or portal integration
Clear escalation behaviour when questions fall outside scope or confidence thresholds
Core Use Cases
This was designed to support high‑volume, low‑risk enquiries, including:
CPD requirements and mandatory learning
Professional and clinical standards
How to make a complaint
Logging CPD and applying for APCs (via links to guides)
Secondary use cases included high‑level registration queries, with links to authoritative sources only.
Risk Management and Safeguards
Given the regulatory context, risk control was a key design principle.
Redirecting users who attempt to provide complaint narratives
Escalating when questions reference unpublished or internal processes
Avoiding summaries or guidance beyond what is explicitly published on the website
Maintaining a friendly but professional tone, aligned with regulatory expectations
Technical Scope
Deployed on the public website via an embedded chatbot widget
No integration with internal systems or databases
Internal testing and client review conducted before launch
No need for an API
Outcomes
The organisation achieved:
A measurable reduction in repetitive enquiries reaching staff inboxes
Faster information discovery for practitioners and public users
Consistent, policy‑aligned responses to common questions
Improved capacity for staff to focus on higher‑value regulatory activities
Importantly, the conservative scope meant the chatbot enhanced service delivery without increasing risk or changing existing processes.
This project succeeded because it:
Treated AI as a supporting channel, not a decision‑maker
Respected the organisation’s regulatory obligations and risk appetite
Started small, with clearly defined boundaries
Grounded every response in approved, published content
Built confidence for future enhancement, if required
