AI gains to soar in the 'second innings' of use

Efficiency gains for generative artificial intelligence may only be in the single digits currently, but that will markedly improve as companies tap into the technology’s reasoning, use AI agents and assess business model changes.

That’s the view of Capgemini Australia chief executive Zhiwei Jiang, who likens the use of generative AI to being in the second innings of a baseball game.

“If you think about what has actually been done in the last two years, the foundation has been built and eventually the Gen AI will become an OS, which is essentially an operating system,” he said in an interview.

“We still sit in the second innings of it, in other words we still do these very specific POCs, the proof-of-concept specific user cases.”

Even as regulators express concerns about Gen AI, including potential bias, Mr Jiang thinks the take-up will start to change from next year as companies draw further on its potential.

“Most clients are still only realising the automation aspect of the AI. We haven’t realised the intelligence, the reasoning side.

“About 5 to 10 per cent are trying to figure out how to use Gen AI to drive business model changes and business growth.

“The real potential has to be how do we create a Gen AI or pure Gen AI business or a new business model? When you go to the third or fourth innings of it, these ideas will come up in 2025 or 2026.”

Australian regulators are, however, taking a cautious approach to AI and how it is employed.

The corporate regulator in October warned that financial services companies were at risk of court action, given many had introduced AI technology without appropriately considering controls or governance.

A survey by the Australian Securities & Investments Commission found nearly half the companies polled did not have policies in place governing fairness or bias in their AI technologies, with even fewer disclosing the use of the programs to consumers.

A Deloitte report released last month identified a lack of consumer confidence that organisations will use AI responsibly was prevalent across all 10 industries examined.

Mr Jiang said he acknowledged guardrails were required with the increased use of Gen AI, but also noted risks to companies that didn’t explore its uses.

“I fully appreciate if you think about Gen AI it does introduce a lot of risk, you can argue there’s hallucination, bias, transparency, there’s quality … getting into the issue of garbage in, garbage out and copyright,” he added.

“Enterprises also need to take a more balanced approach. There’s also a risk of potentially missing all the benefits that AI and Gen AI will actually get you.”

Mr Jiang said a “cost-conscious” mindset among executives was constraining technology investment.

Banks around the world are using and testing AI and Gen AI uses within their front-end businesses and back-office functions.

Commonwealth Bank chief executive Matt Comyn used a briefing last week to outline the bank’s ambitions and current use cases for AI and Gen AI.

He said CBA’s AI-powered technologies had slashed call centre wait times by 40 per cent, while – in combination with other systems – AI helped drive a halving in scam losses.

CBA is also putting guardrails in place to avoid any problems AI may cause and to monitor outcomes.

“At the moment we have got a human in the loop in these Generative AI experiences,” Mr Comyn said last week. “You can imagine this is an area, as it should be, that we would put a very high bar on accuracy.”

Among Capgemini’s top five technology trends for 2025 is increased takeup of reasoning AI agents, which will start operating more autonomously as well as providing more reliable outputs. AI agents are independent systems that conduct tasks without human involvement.

Another anticipated trend relates to AI-driven robotics and a blurring of lines between humans and machines, as robots take on more complex roles in industries including manufacturing, logistics and agriculture.

Mr Jiang, appointed to the role in February, joined Capgemini more than a decade ago after stints with Deutsche Bank, Goldman Sachs and Morgan Stanley.

Capgemini does not disclose who its customers are, but Mr Jiang said it worked with most of Australia’s top tier banks and also had customers in the share trading industry and public sector.

Morgan Stanley’s global chief executive Ted Pick in June told a conference the use of AI for its financial advisers was potentially “game-changing”, and could save them between 10 and 15 hours per week.

Commenting on how financial services firms were implementing AI more broadly in private wealth, Mr Jiang said: “Gen AI has the power to automate, to scale, to personalise. So, this is a perfect user case to augment, I don’t really see people actually using Gen AI to replace these FAS (financial advisers).

He noted advisers could use the technology to summarise relevant research reports and scan sharemarket filings for insights, to assist them in personalising a recommendation.

Separately, Mr Jiang also highlighted that the AI sector was grappling with a shortage of employees with the requisite experience and knowledge, given those that worked within the industry required a host of different skills.

Mr Jiang said policymakers and universities in parts of the US and in Singapore and China were ahead of other markets in trying to address this problem.

It’s a point that CBA is looking to address, with the bank recently appointing its first AI scientists who will, alongside engineers, tackle complex technology problems.

“Through building with AI, we are upskilling all of our engineers, empowering them with skills, tools and infrastructure they need in order to build Gen AI-enabled solutions and working with AI,” CBA CIO Vicky Ledda said.

This article first appeared in The Australian.