One NZ launches 'Network Concierge' to speed up network issue diagnosis for support teams
2 December 2025
One New Zealand has rolled out a new AI agent called Network Concierge, built on Amazon Bedrock by AWS. It pulls data from multiple internal systems into one place so teams can quickly diagnose coverage-related customer issues. Early results show an 80% reduction in analysis time, which means fewer transfers, clearer updates, and faster resolutions for customers.
What this means for customer support
- Single place to check a customer's local network health - no more jumping between tools.
- 80% faster analysis time, so reps can move from "let me check" to clear next steps in minutes.
- Summarised context: performance data, alarms, and past incidents in a plain-English view.
- Better first-contact resolution and more confident conversations during outages.
Here's how it works in practice: a rep enters the customer's details and asks the agent to review the local network. The agent pulls performance signals, recent alarms, historical incidents, and site notes, then returns a short summary with guidance on what to do next. Support can give the customer a straight answer without waiting on a long back-and-forth with engineering.
"Building on the success we've seen with AWS's generative AI in our contact centres, we wanted to apply this technology to speed up diagnosis and resolution when our customers report network and coverage related issues. Our Network Concierge agent pulls everything into one place and gives our engineers clear answers in minutes, which in turn reduces wait times for customers," says Thaigan Govender, GM Mobile Networks, One NZ.
Faster generator deployment during power outages
During storms and extended outages, teams need to prioritise mobile generators fast. That used to mean manually trawling through system data. Now, with Amazon Bedrock bringing the data together, the AI agent identifies sites low on backup power, flags which ones are most critical, and recommends where generators should go first. This approach was tested during recent severe weather in the South Island and helped teams respond faster.
Why this matters for your queue
Less time spent digging through tools means shorter handle times and fewer escalations. Reps can stay present with the customer, explain what's happening, set expectations, and move to action. AI takes on the repetitive data gathering; people focus on the high-value conversation.
Building a stronger, more reliable network for Aotearoa
"These new tools let our teams work smarter and much faster," says Thaigan. "Our goal is to become the world's most AI-enabled telco. Working with companies like AWS helps us get there. By weaving AI into the way we run the network and support customers, we're changing how we respond to outages, solve issues and deploy people on the ground.
"AI handles the repetitive, data gathering type work that is time consuming, and our people focus on the tricky and important things. AI-first, human when it matters most."
One NZ is now exploring deeper integration between AWS agentic AI and its existing systems to automate more of its incident management and field-force workflows.
Manuel Bohnet, Country Manager for AWS New Zealand says: "Today, losing network connectivity even for a few minutes can disrupt people's daily activities, and that's why we think One NZ's Network Concierge is so impactful. By reducing analysis time by 80%, it can help Kiwis get back online much faster. Looking ahead, this technology will predict issues before they occur and automate repairs in the background before they turn into major disruptions. One NZ's implementation of Amazon Bedrock and agentic AI demonstrates how leading New Zealand companies are leaning into the latest AI technologies to improve customer experience and operational efficiencies, and we look forward to supporting their AI journey."
Practical next steps for support leaders
- Map your top network-related contact drivers and define clear prompts the agent should answer.
- Connect the agent to your CRM/case system so summaries and next steps log automatically.
- Create simple playbooks for outages: what to say, what to check, and when to escalate.
- Track handle time, first-contact resolution, and customer effort score before and after rollout.
To learn more about the underlying tech, see Amazon Bedrock from AWS here. If you're upskilling your team on practical AI for customer-facing work, explore role-based programs here.
For real-time coverage updates, head to one.nz/coverage.
Your membership also unlocks:
