The real world of artificial intelligence in New Zealand

A recent survey of executives in New Zealand and the Asia Pacific region revealed some important insights on how Kiwi organisations view AI and how it will challenge the competitive landscape over the coming years.

Surprisingly, there was quite a difference in sentiment towards AI in New Zealand compared to our close neighbours in Australia, with NZ execs seeing the AI need as more urgent.

‍

In addition, 85% of New Zealand organisations see AI as a positive technological advancement while just 13% see it as a threat to their jobs.

Where to begin?

Most NZ organisations believe AI will help them increase process automation and employee productivity, while AI embedded into processes such as customer service will help uncover new insights, such as product and service recommendations or issue and root cause discovery. As always, good useable AI depends on the quality of the data used to train the models.

Barriers and solutions to AI adoption

High costs, governance and regulatory implications were seen as high barriers to AI adoption.

AI technology is progressing quickly and chatbot solutions, like FAQ Bot, are low cost and low risk AI solutions with the potential for high return. In New Zealand, 48% of organisations surveyed by IDC say they are currently evaluating chatbot technology while 12% say they already have or are currently implementing the technology.

At the other end of the barrier scale simple things like not knowing where to start or not understanding the potential are holding organisations back.

Why do AI projects fail?

Unsurprisingly, every failure is different. Louise Francis outlined the top reasons for failed AI projects. In the Top 5 were

• Solving the wrong problem, solving a problem that didn’t need solving, or simply using the wrong tools
• Taking on a project that is too big or too grand
• Poor data quality (includes ongoing data governance and cleansing)
• Trying to automate everything when AI models may not be ready for it
• Forgetting to regularly monitor, update and adjust the solution after the initial project is complete

Ethics in AI

There  many ethical challenges when it comes to AI – from algorithmic bias, transparency, surveillance and fraud to ‘. There’s a range of tools to address these concerns, says Francis, and some are simple while others are not so simple. Examples include ensuring data diversity, compliance, transparency, using algorithmic impact assessment tools, and being prepared to just pull the plug.

Your AI roadmap – top tips from Louise Francis

Tip one: impact

When thinking about implementing AI it’s important to consider how it will affect how your organisation and, in some cases, your customers. Identify the right use cases and even more importantly make sure it’s aligned with how your customers want to use AI. Remember to start simple and ensure it will make a difference.

Tip two: people

Talent is key so think about how you could or work with the right partner along the journey. And don’t forget the plan to deal with employee change management.

Tip three: wider context

Lastly, don’t forget to consider ethics, compliance and regulation when implementing AI.