Polyaire deploys NeonNow to lift contact centre insight

Polyaire has deployed NeonNow's CX Insights system to analyse customer calls across its contact centre, linking call data from Amazon Connect with customer records in HubSpot.

The Adelaide HVAC manufacturer activated the system on 14 September 2026 and said it delivered changes within 24 days, including faster response times and lower call abandonment. The project is part of Resolve, a three-phase programme to build an end-to-end customer engagement centre.

Polyaire, which makes HVAC equipment and the AirTouch smart home air-conditioning control system, uses its customer engagement team, known as the Airmates, to support homeowners, installers and trade customers across wholesale, branch and digital channels. About two years earlier, the business moved its contact centre to Amazon Connect as the programme's first phase.

The next stage depended on understanding why customers were calling. That had been difficult because call data and customer records were stored in separate systems, leaving staff to download data, listen to recordings and manually match calls to HubSpot records.

This left management without a consistent evidence base for recurring issues in customer conversations, including onboarding, product support and lead generation. Polyaire assessed potential partners before selecting NeonNow for the work.

"The NeonNow team asked the right questions and listened until they understood exactly what we needed to learn from our calls," said Martin Mueller-Weichert, Head of Customer Experience, Polyaire.

NeonNow connected Amazon Connect call data with HubSpot customer records in provisioned data lakes. Its CX platform then used Amazon Quick to identify recurring themes and customer sentiment across calls. Polyaire also introduced three custom dashboards, including live agent wallboards, and moved agents to a single workspace containing both calls and customer records.

The setup also included a Model Context Protocol server built on Amazon Bedrock AgentCore, allowing executives to query the data in plain language through Claude Desktop. NeonNow also reviewed Polyaire's playbook, key performance indicators and service-level agreements, and made recommendations on operating practice.

Operational changes

Between 14 September and 7 October, the contact centre handled 953 contacts across five queues. During that period, AI completed 971 quality evaluations without supervisors manually reviewing calls, recording an average score of 92.92%.

Polyaire said average speed to answer for inbound calls improved by 36%, while inbound queue wait times fell by 33%. Average handle time fell by 19% across all queues, and call abandonment dropped by 32% to 2.76%.

The company can now track customer sentiment on every analysed call, which had not previously been possible.

"For the first time we could conduct detailed per-queue agent evaluations for 100% of calls," said Mueller-Weichert.

Early warning

An early test of the system came when connectivity issues rose from about 10% to 45% of initial cases during the pilot period. Using AI logging and testing tools in the platform, the teams traced most of the increase to unstructured product reference material and confusion during user adoption.

Polyaire updated knowledge-base articles and expanded the query catalogues available to frontline agents. The issue was resolved before it developed into a larger escalation.

Executives now have a clearer picture of what customers are reporting through the contact centre, while team leaders can monitor performance in real time. For Polyaire, which receives support calls after products have been installed or purchased, contact centre data has become a direct source of information on product issues, installer experience and homeowner needs.

"It's not just about what happens before the sale. It's after the sale that customers truly remember you," said Mike Powrie, Chief Executive Officer, NeonNow.