Eight steps to avoid breaking the bank on AI's seductive promise
“I don’t know where AI is going, but we have to spend or else we could miss out.”
Far too often, I’m hearing this exact line from CIOs who in any other context would be much more circumspect when considering their tech spend. But not in the age of AI. Instead, I’m hearing panic, uncertainty and urgency to do something even if it could be the wrong thing. And, frankly, it could be the wrong thing.
I’ve been struck by some of the stunning efficiencies we’ve seen AI drive in advanced manufacturing. But when Gartner predicts that 30% of generative AI projects will be abandoned before the end of 2025, all of us should pause and consider the right path forward. After all, that’s a full third of the AI projects being undertaken now, and it doesn’t include those that have a sunk cost fallacy kick in. There are going to be plenty of business leaders who simply can’t bring themselves to hit the kill switch on something that is failing to return value after betting big on it — which will only lead to bigger losses and future entrenched inefficiencies.
But this concerning trend towards doing something, anything, even the potentially wrong thing, is driven by several factors. One factor is the macroeconomic situation: many believe Australia has lost its way with quality of life, cost of living, and general sentiment. This is causing business leaders to feel like they have to do something or be trapped in that same spiral of negativity.
Secondly, there’s the advent of AI and what it promises. AI’s seductive promise and power make company leaders feel like they have to invest heavily in technology to keep pace. The power is truly there, but its application in many business cases is far from clear. The expectation is that it will help them to stay competitive and they can use it to transform the business — but how? That part is fuzzy.
The first step for anyone considering AI for the business is getting your house in order. One of the core mistakes that we see companies make over and over again is that they overcommit to AI without first addressing critical infrastructure needs. Many businesses leap into AI projects, imagining breakthrough insights and automation, but overlook the essential groundwork: data consolidation, system integration and platform optimisation.
Companies need to be flexible and embrace changes in the marketplace. But this is inherently risky and needs to be planned carefully.
A manufacturing company recently decided to transition from producing metals to health and beauty products. In the process of doing that, the company has subsequently invested millions in platforms like Salesforce and ERP systems, hoping to revamp its operations. However, the sheer scale of this spend is disproportionate to its financial performance — there’s no way what it is spending on IT systems is justified, given its market cap and manufacturing capabilities.
The continued spending, despite limited potential returns, is a common gamble that we see time and again: investing heavily in AI, because it is the thing to do, without a clear strategy for how it will produce ROI.
This pattern is symptomatic of a broader issue in Australian business. It is all too common to throw money at a vague goal of business transformation without establishing the necessary infrastructure or business case to do so.
Recently, Australia’s AirTrunk was acquired for $24 billion. This is the fifth biggest acquisition deal in ASX history, and the largest data centre deal globally. The reason for it is clear: there is demand for data centres in a world of AI, and AirTrunk certainly has massive capacity already built. Nonetheless, it is going to be interesting to see whether the returns will justify that price tag.
For other organisations, it’s not necessary to invest $24 billion to step into AI now. Indeed, AI neither needs to be — nor should be — a matter of simply finding money to make investments. Organisations should focus on building an action plan first, and then invest strategically to fulfil the needs of that action plan. For most organisations, that action plan should look something like the following:
- Ensure your data is clean, free from conflicts, and consistent across all departments to avoid errors when deploying AI models. This is essential for accurate AI-driven insights.
- Consider what your ideal AI-enabled infrastructure would look like without existing constraints. This helps identify transformative changes and stay competitive against new market entrants.
- Prioritise AI applications that directly support key business objectives, such as improving product quality or reducing waste. Avoid implementing AI solely for its novelty.
- Form an AI governance group to define policies and procedures around AI usage, addressing data security, ethical considerations and responsible practices.
- Incorporate AI tools that match the intuitive digital experiences employees have outside of work to prevent staff from using unsanctioned tools, which could compromise data security.
- While future AI developments are fast-paced, continue to align AI initiatives with core goals such as revenue growth, sustainability and risk reduction to ensure resilience.
- Work closely with technology providers to stay informed about available resources, training and AI-related support. For instance, Microsoft’s AI master classes can provide deeper knowledge and tools.
- Utilise industry 4.0 or other digital maturity assessments to understand where your organisation stands in terms of people, processes and technology, and to establish a roadmap for AI integration.
Following a methodical and clearly mapped-out action plan like this is the most effective way to ensure that the investments that you are making into AI are not ‘expenditure for the sake of it’, or a ‘Hail Mary’ hoping to latch onto AI to mitigate challenges that you’re facing in the market. Instead, it will ensure that at each point in time, the journey towards AI is also returning value to the business.
What has been firmly proven is that the investments into AI that have a positive return have one thing in common: they are the result of establishing solid foundations and successfully navigating the hype.
Despite years of explosive data growth, there may not be enough for AI
Enterprises have reached a fork in the road, where they must either find more data or shrink the...
Is the Australian tech skills gap a myth?
As Australia navigates this shift towards a skills-based economy, addressing the learning gap...
How 'pre-mortem' analysis can support successful IT deployments
As IT projects become more complex, the adoption of pre-mortem analysis should be a standard...