Overview
Arinco partnered with Tauranga City Council to reduce LGOIMA email triage from up to 5 hours to as little as 10 minutes by introducing an AI-assisted solution into its review process. To better meet citizen expectations and improve the speed and consistency of handling official information requests, the council needed a solution that could streamline review processes without requiring a major workflow redesign. Arinco delivered a targeted AI-assisted solution using Microsoft Graph and Azure AI, focused on streamlining early-stage email triage while maintaining traceability and human oversight. This approach delivered immediate operational value in a production environment while establishing a scalable foundation for future AI-enabled initiatives.
About Tauranga City Council
Tauranga City Council is the local authority for one of New Zealand’s fastest-growing cities, delivering a wide range of civic, infrastructure, regulatory and community services. The council supports a diverse and evolving population and continues to invest in digital capability to improve service delivery and operational efficiency.
Business challenge
Tauranga City Council is responsible for responding to requests made under the Local Government Official Information and Meetings Act 1987 (LGOIMA), which gives members of the public the right to access official information held by the council. These requests can cover a wide range of material, including emails, reports, decisions and internal correspondence, and must be responded to within strict statutory timeframes.
Handling these requests requires more than simply locating documents, with email review and triage often taking staff up to five hours per request. Staff must work across large volumes of information to identify what is relevant, remove duplicates and noise, preserve context, and ensure responses are accurate, traceable and defensible, with a clear link back to source material.
LGOIMA requests regularly involve reviewing anywhere from around 100 to more than 3,000–4,000 emails per case, with approximately 1,000 emails being typical for many requests. Council teams also process hundreds, and in some cases thousands, of LGOIMA requests every year, creating significant operational overhead and time pressure.
This is made more challenging by how information is stored across the organisation. Relevant content can be spread across email systems, Teams, OneDrive, SharePoint and other repositories, making it difficult to review material consistently and efficiently.
The most time-intensive stage was the email review process. Staff often worked through thousands of emails to identify a much smaller subset of relevant content. In one case, more than 2,200 emails were reviewed to identify just 31 items that were ultimately in scope. Council stakeholders estimated that processing an initial set of 200 emails could take approximately one hour of manual effort, highlighting the scale of effort required for larger and more complex requests.
The council had already been taking a practical and forward-looking approach to AI, creating an opportunity to explore how it could be applied to improve LGOIMA handling while maintaining the rigour required for official responses.
Solution
Arinco delivered an AI-assisted LGOIMA solution for Tauranga City Council, building on the council’s existing focus on practical AI adoption across its operations.
The engagement focused on a high-impact use case where AI could deliver immediate value. Email triage was selected as the most time-intensive and operationally constrained stage of the process, with a clear opportunity to reduce manual effort while maintaining governance, traceability and control.
The solution was specifically designed to help reviewers work through very large email datasets more efficiently, including cases involving thousands of messages and substantial amounts of duplicated or low-value content.
Targeted triage approach
The solution was designed to improve efficiency and consistency while aligning with existing processes. It enables staff to:
- Remove duplicate and irrelevant emails
- Isolate requester-related correspondence
- Separate administrative content from substantive material
- Narrow large datasets into a smaller, relevant working set
- Surface evidence for specific case questions
This ensured the solution delivered immediate value while helping staff work more efficiently and with greater confidence, without disrupting established processes and governance frameworks.
AI-enabled analysis and traceability
Built using Microsoft Graph and Azure AI, Arinco’s solution introduced capabilities to support both efficiency and defensibility. This included:
- Deduplication and clustering of email content
- Identification of related or similar information
- Evidence-backed responses with traceability to source emails
- Full activity history to support audit requirements
Importantly, the solution was designed to support, not replace, human decision-making, ensuring responses remained traceable, defensible and grounded in source material. Reviewers remain responsible for assessment, redaction and release decisions, while the AI-assisted capability reduces the volume of material requiring manual review and preserves context throughout the process.
By focusing on a targeted, high-value use case and leveraging capabilities from Arinco’s AI Build Centre, Tauranga City Council moved from proof of concept to production quickly. This delivered immediate operational improvements while establishing a scalable foundation for future AI capabilities across Microsoft 365, Azure AI and Microsoft Foundry.
Outcomes
The AI-assisted LGOIMA solution is now in production and delivering measurable value in Tauranga City Council’s correspondence triage workflow. It has significantly reduced the time and effort required to process requests involving hundreds to thousands of emails, while improving consistency, traceability and confidence through a human-led review approach.
Key outcomes include:
- Significant time savings: Reduced email triage from up to 5 hours to as little as 10 minutes, with complex cases (20+ hours) also reduced to a fraction of that time.
- Faster evidence retrieval: Reduced follow-up evidence-finding from up to 60 minutes to near-immediate access to relevant supporting material, accelerating response preparation.
- Improved defensibility: Enabled evidence-backed responses with clear traceability to source emails, supporting accurate and justifiable responses.
- Enhanced auditability: Provided full visibility into review activity, including actions taken and how decisions were reached.
- Seamless adoption: Integrated into existing tools and review processes without requiring extensive retraining or major redesign, supporting review workflows involving hundreds to thousands of emails per request.
- Scalable review capability: Enabled council teams to process requests involving anywhere from around 100 to more than 3,000–4,000 emails more efficiently, significantly reducing administrative effort across high-volume LGOIMA workloads.
These outcomes demonstrate a reduction in time and effort for staff, along with improved quality, consistency and trust in LGOIMA responses. Tauranga City Council now has a proven example of AI delivering operational value while maintaining strong governance and human oversight.
What's next?
With AI-assisted email triage now in production, Tauranga City Council is positioned to extend these capabilities across additional information sources and stages of the LGOIMA process.
The next phase will expand coverage beyond email to platforms such as SharePoint, Teams and other records repositories, enabling a more consistent, end-to-end approach. Planned enhancements include redaction support to help identify and manage sensitive content more efficiently, while keeping final decisions with reviewers.
There is also an opportunity to integrate more of the workflow into a coordinated AI-assisted process, embedding these capabilities more deeply into day-to-day operations.
Building on early success with Microsoft 365 Copilot and Copilot Studio, the council can further leverage Microsoft Foundry and Azure AI to support ongoing innovation, with continued support from Arinco.
