Making the Most of Your Data: AI-Enabled Solutions with UC & IEEE

On Thursday, October 16, Business Canterbury, in collaboration with the University of Canterbury and IEEE, hosted a half-day event titled “Make the Most of Your Data: AI-enabled Solutions.” The session brought together data experts, academics, and business leaders to explore how AI and data infrastructure can be harnessed to drive productivity and profitability.

From Hype to Reality: The Current State of AI

Dr. Elsamari Botha, Associate Professor at UC opened the event with a candid look at the current landscape of AI adoption. She described the industry as moving beyond the initial excitement into what she called the “post-enthusiasm wave.” Organisations are feeling pressure to adopt AI, but many are encountering significant challenges.

A striking statistic from MIT revealed that 95% of AI projects are failing, a trend mirrored in New Zealand. Companies that rushed to replace human roles with AI—such as Klarna—are now reversing course, realising that technology alone cannot replace strategic thinking. Elsamari explained that AI is currently in the “trough of disillusionment,” where expectations have outpaced reality. She introduced the concept of a “jagged technological frontier,” where AI performs brilliantly in some areas but poorly in others.

The key message: AI should not be adopted for its novelty. Instead, it must be aligned with clear business objectives to succeed.

Laying the Groundwork: Data Infrastructure Matters

Dr Rakesh Kumar, President, Solid-State Circuits Society, IEEE followed with a deep dive into the importance of data infrastructure. He emphasised that high-quality data is the foundation of any successful AI initiative. While many organisations are collecting data, few know how to assess its quality or use it effectively.

Rakesh introduced the FAIR principles—Findable, Accessible, Interoperable, and Reusable—as essential guidelines for managing data. He also highlighted common challenges, including privacy concerns, regulatory compliance, and the disconnect between domain experts and data scientists.

Discussions also touched on practical considerations such as cloud vs. on-premises storage, data ownership, and latency issues for large spatial datasets. These decisions can significantly impact the usability and security of data in AI applications.

AI in Action: A University Case Study

One of the most compelling parts of the event was a case study presented by Paul Benden, Lead AI & Data Scientist at UC, Faced with the challenge of interpreting over 300,000 words of complex academic regulations, the university developed an AI-powered solution to streamline graduation eligibility checks.

The project combined generative AI (Google’s Gemini Pro 2.5) with traditional AI (Prolog) to create a web application that could reliably process student records against university rules. The process involved extracting text from PDFs, converting it into structured data, and applying logic rules to determine eligibility.

With a modest investment of $3,000 in AI credits and 40 hours of development time, the university created a solution that saved significant time and reduced errors—demonstrating the power of hybrid AI approaches.

Strategic Implementation: Start with the Problem

The panel discussion reinforced the importance of a strategic approach to AI. Rather than starting with the technology, businesses should begin by identifying the problem they want to solve. Blanket licensing of AI tools without clear use cases was discouraged.

Elsa Marie encouraged organisations to think beyond efficiency gains and explore opportunity-driven AI—solutions that enhance offerings or create entirely new revenue streams. A hybrid approach, using different AI models for different tasks, was recommended as a practical way forward.

Challenges and Considerations

The panel didn’t shy away from discussing the challenges of AI adoption. Poor data quality, lack of AI literacy, and unclear return on investment were cited as common barriers. There were also concerns about the environmental impact of AI, the proliferation of uncoordinated tools (referred to as the “Excel moment”), and academic integrity issues as students increasingly use AI tools.

Another emerging concern was the degradation of AI-generated content over time, as models begin to train on hallucinated or low-quality data.

Final Thoughts

This event underscored that while AI holds immense promise, its success hinges on strategic planning, robust data infrastructure, and a clear understanding of its capabilities and limitations. Businesses must move beyond the hype and focus on purposeful, well-governed AI adoption to truly unlock its potential.

What’s Next for Businesses?

1. To help businesses move forward, the panel offered several actionable recommendations:

• Evaluate current data infrastructure and identify gaps
• Define clear business problems before implementing AI
• Explore platforms like IEEE Dataport for data sharing and competitions
• Consider hybrid AI approaches tailored to specific use cases
• Establish governance frameworks with human oversight
• Upskill staff to work effectively with AI tools
  
  

2. Business Canterbury also announced upcoming workshops:

• Join the Power Up Your Leadership Workshop | 29 October
  Led by Amy Jones, this workshop helps SME leaders build influence and lead with clarity.
  View details and register →
• Register for the AI in Action Workshop | 5 November
  A hands-on session designed to help businesses embed AI into everyday work.
  Learn more and register here →
  
  

3. Partnership opportunities - 

The University of Canterbury is offering partnership opportunities for businesses interested in exploring AI applications, while IEEE Dataport provides a collaborative space for data-driven innovation.