AI won't transform your business unless you transform it first

Organisations need to overhaul their processes to truly reap the benefits of artificial intelligence, according to experts.

Bolting the technology to existing systems will not bring about the drastic changes that AI providers are touting.

“It's a huge change and that's why the dynamic is not going to be as fast as people imagine,” said Denis Torii, vice president analyst at Gartner.

“It's not an IT story for sure. It's not [the case of] only the business stakeholders who are pushing us to adopt more AI,” he said. "It's actually an organisational story, where the organisation needs to adapt to it."

AI is only working as an accessory today.

“Right now [organisations] are doing the same stuff but with AI in the middle, and that probably is not going to bring that transformation that people are expecting,” he told The National on the sidelines of the recent Oracle AI World event in Las Vegas.

Currently, organisations are using AI across the enterprise with a range of tools – from chatbots and software coding agents to agentic systems capable of acting autonomously across workflows, a McKinsey report in August found.

However, enterprise-level financial impact remains limited. Only 37 per cent of respondents reported an AI contribution to their organisations' earnings before interest and tax (Ebit), unchanged from a year earlier, McKinsey's Global Survey on the State of AI found.

01:55

A crisis will spark rush to regulate AI, says Apollo's Marc Rowan

To benefit from AI requires a shift in leadership mindset, according to Gabby Fredkin, head of analytics and insights at Adapt, a specialist IT research advisory.

A fundamental change is needed to go from “we're just going to layer technology on, to we're actually going to think about the ethos of our organisation, what we do, the type of roles that we have”, he said.

“So, it's a complicated process.”

The organisations that generally succeed with AI tend to have a leadership that has a good conceptual understanding about it and encourages people to innovate, he added.

The experience of a small group of high performers highlights the technology’s successful use, the McKinsey report said.

“High performers pursue growth and/or innovation alongside efficiency; they fundamentally redesign workflows that are enabled by AI rather than insert AI into existing ones; and they back their deployments with the leadership commitment and operational rigor needed to achieve real gains," the report found.

Mr Fredkin said companies must look into their operating models to see how they can go beyond traditional processes and capture new revenue streams.

“If you're a bank and you've got data because you've got investments in different infrastructure, do you want to get into infrastructure as an investment? Do you want to start a real estate investment trust? How are you thinking about what you have and what you do, and what are the parallels outside of that?", he said. Such a shift would be harder for some organisations than others, he added.

Cost factor

AI-related operating costs are beginning to constrain its use for about one in five organisations, the McKinsey report said, even as most respondents expect their companies to increase AI investments in the year ahead.

01:28

Sam Altman says humans must remain in control of AI

Companies are taking a cautious approach as they prepare their data and processes for AI integration, said analyst John Reed, co-founder of US-based tech platform diginomica.

“If companies know that it's not going to be that expensive to use these technologies, I think they are going to find significant benefits,” he said.

AI could prove particularly valuable in complex, mission-critical tasks that humans struggle to perform in real time, Mr Reed added.

“For example, it's very difficult for a human to assess global supply chain impacts in real time … these [AI] systems actually are able to, if you provide them with the right data.”

However, adoption will take time, as many companies grapple with fragmented data and legacy systems, he added.

The trillion-dollar ROI challenge

Investment in AI is soaring. Worldwide spending is forecast to total $2.7 trillion this year, up 49.5 per cent annually, Gartner said in a report in September.

AI infrastructure, particularly data centres and specialised servers, remains the biggest spending driver, with Gartner describing the buildout as the “largest infrastructure project humanity has ever undertaken”.

Goldman Sachs Research separately estimates $1 trillion in global AI-related investment this year, including $581 billion in the US. It expects AI capital expenditure to rise from 0.9 per cent of global GDP in 2026 to 1.4 per cent by 2028.

But the spending boom is raising questions about the revenue needed to justify it. The AI industry needs to earn $6 trillion in annual revenue by 2031 to justify the capital being deployed, a report by consultancy Bain in September found.

“Return on investment is still one of the biggest issues [with AI],” said Mr Torii. “People are putting all this faith in AI, [with the assumption that] there's a lot of potential.

“When you look at the broad picture, it's still lacking evidence that return on investment is there,” he said, adding, however, that vendors need to invest heavily upfront.

“They need to help customers somehow to get to that journey. So there's a lot of friction and inertia that needs to be overcome.”

He compared AI's trajectory to the early days of cloud computing, when heavy infrastructure investment preceded widespread adoption. The question, Mr Torii said, is how long it will take for AI to bridge the gap.

Measuring returns is another challenge, Mr Fredkin said. “They [organisations] feel a productivity uplift. They feel better employee engagement. Putting an actual figure behind that is incredibly complex,” he said.

“From the perspective of the CFO [chief financial officer], the calculations for how they look at ROI [return on investment] have changed significantly. It's not a line item any more; it's a meter.”

He cited an emerging term “double T Ebitda”, which refers to earnings before interest, tax, tokens, depreciation, and amortisation.

Unlike blockchain and metaverse technologies, AI is already gaining traction across businesses and is likely to see wider adoption, Mr Reed said.

Looking ahead, he is slightly pessimistic about its broader implications. “I think we're in some significant trouble because we put a lot of eggs in the basket of a technology that, while it's good for business, I'm not convinced [about] the ultimate impact in terms of surveillance and disinformation and deep fakes and the inability to trust content and the displacement of humans from jobs,” he said.

“But one of the things about technology is that it can be decentralised and changed and appropriated too. And there's a lot of very thoughtful people that are using some of these tools for really cool, worthwhile projects as well,” he added.