AI is having its ‘cloud’ moment: how can you make the right decisions for your business?

Ten years ago, many Australian businesses embarked on a cloud strategy, the idea being to move tech spending from capex to opex, and to become more agile and able to keep up with the rapid technological pace of change.

Today, just over two years since artificial intelligence burst into the public consciousness with the launch of ChatGPT, AI is having its own ‘cloud’ moment, and many Australian businesses are about to embark on their own AI journey. As with cloud, the goal is to become more efficient, offer better customer service and keep up with technological progress. 

The burning questions for many Australian businesses, however, is how to create an AI program, what are some of the challenges and mistakes, and how do you measure its success?

AI is top-of-mind for many Australian organisations

Local businesses haven’t been slow when it comes to using AI, with a recent CSIRO report finding 68 per cent of Australian companies have already implemented AI in some form, and a further 23 per cent are planning to do so within the year.

And those local companies are already reaping the benefits of their early AI adoption. In terms of improved customer service, the same CSIRO report found domestic organisations are earning an additional $500,000 in revenue from customer service bots which improve the customer experience through round-the-clock support, enhanced response times and better, more personalised product recommendations.

AI is also relieving administrative burdens, improving productivity, and saving time. In fact, the CSIRO found the businesses it surveyed reported time savings of 30 per cent across all AI-related initiatives. Additionally, AI is a tide that’s lifting all boats, with Goldman Sachs reporting AI adoption could lift global productivity growth by 1.5 per cent over ten years and drive a seven per cent increase in global GDP.

Building an effective AI strategy

For Australian businesses contemplating an AI journey there are several steps they need to take to make it a success. According to a Stanford University study, there are four factors organisations need to consider when beginning an AI program. 

The first consideration is to ask yourself what the business opportunity or problem is that AI can potentially solve. AI has incredible potential, but you won’t realise that potential simply using AI for the sake of it. “Keeping up with the cool kids is not the basis for a solid business strategy.”

The next factor is to think about your timeline. AI is moving fast, but the investments needed, both financially and in terms of resources, are substantial and your company needs to be able to sustain the initiative.

Wait too long, and you could fall behind, but rushing into something without adequate preparation is equally fraught and can lead to failure or an outcome that’s simply not up to scratch.

Third, plan, and plan some more. Create a roadmap to get you where you want to go with AI. Your plan should lay out the steps for implementation, the necessary support and the KPIs you’ll use to evaluate success.

Finally, you’ll need strategy around data (is your corporate data of high enough quality?) as well as around the AI model you’ll use and the infrastructure you’ll run your AI model on. 

Common mistakes when creating an AI strategy

Futurist Bernard Marr observes one of the biggest mistakes companies make when creating an AI strategy is not having clear objectives about what they want from AI. While some companies are early adopters, they often fail to define what they want their AI initiative to achieve.

Those same organisations may not have a change management strategy in place allowing for a shift in company culture brought on by AI, and they might also overestimate what AI can do for them. AI is powerful, but it requires a lot of input and management if it’s going to be effective. 

Organisations can also neglect to find the talent needed to run their AI program – something that’s tough in Australia with the IT skills shortage – and they may also overlook having a data strategy in place. Data is the lifeblood of any AI system, but it must be clean and well-organised if it’s going to be useful. 

Finally, it’s easy to underestimate the budget and resourcing required for AI, and businesses must be mindful AI is an ongoing project, not just a one-off that can be ‘set and forget.’ Without adequate resourcing, the AI initiative is on a path to failure.

Measuring success

How will you know if your AI program is a success? It comes down to user satisfaction, and the impact it makes.

You need to measure how easy it is for people to use your AI, particularly if it’s a chatbot or some other customer-facing tool. This means it’s important to collect feedback to understand user pain points, and then use the feedback as data to reinforce the AI’s training. 

Measuring success also comes down to evaluating your system. This could be through using historical or simulated data to test performance and usability, or live data with real users. A third option is a hybrid of both.

Many Australian companies are embarking on an AI journey, but success comes down to planning, having realistic expectations and adequately resourcing your AI program. AI will massively change the way your organisation does business and how it serves its customers, but understanding its limitations and having clear goals are essential.