Australian organisations are keeping pace with global innovation in generative AI, according to IBM Consulting Global Head of Generative AI Matthew Candy. A local legislative focus on regulating high-risk use cases could also foster AI’s potential in the local market, he said.
Candy recently visited Australia and the broader APAC region to meet some of IBM Consulting’s regional clients and partners, many of whom will be moving from piloting generative AI to implementing models at scale throughout the calendar year 2024.
Speaking with TechRepublic Australia, Candy predicted a move to smaller AI models and the emergence of new digital products and services. He suggested organisations looking at scale needed to focus on strategy and business value as well as aspects like governance.
How does Australia stack up with the world on generative AI?
IBM’s Global AI Adoption Index 2023 pegged Australia as a lagging market for AI adoption. It found only 29% of Australian organisations were actively deploying AI in November of 2023, well behind early adopters India (59%), China (50%), Singapore (53%) and the UAE (58%).
SEE: Australian small and midsize businesses are at risk of getting left behind on AI
However, Candy said there is evidence Australia is embracing generative AI. He said even organisations in traditional or regulated industries, as well as government agencies, are moving forward, and the local market showed a clear understanding of where value could be derived.
“There is some pretty innovative work being done in Australia,” Candy said. “One of the teams I’ve spent some time with are building new innovative digital products and services powered by generative AI models; I definitely don’t see anyone behind just because they are in Australia.”
Generative AI use cases being rolled out now in Australia
IBM Consulting has seen a lot of product pilot work in Australia, with a growing number of these projects now active beyond the pilot and experimentation phase that characterised 2023. From his time spent in the Australian market, Candy said some interesting client use cases include:
- Companies in the asset-intensive utilities industry are using generative AI-powered assistants to help executives make investment decisions on asset management and managing an asset portfolio, which could result in significant increase in savings.
- Utilities companies are wrapping generative AI assistants around complex knowledge bases, like standard operating procedures, to allow the likes of network controllers to chat with complex document sets to take the friction out of some tasks.
- Universities are using generative AI to help generate more personalised content to support student communications while enabling students to interact in a conversational manner with the course content they are learning.
- The end-to-end software development life cycle is being improved in at least one big bank, where IBM Consulting is supporting the use of generative AI to do things like translate project requirements into creative outputs such as user stories and code.
- An Australian government agency is using large language models to create a brand new skills training platform.
‘Value pools’ and people among considerations when scaling AI
Candy was appointed to lead IBM Consulting’s 160,000 global consultants into the generative AI age in August 2023. He said Australian organisations looking to scale generative AI experimentation into implementation this year should keep a few things front of mind.
Have a clear vision and strategy
The foundation for generative AI success is having the right kind of strategy. This will be based on the identification of a “North Star,” or articulation of a vision for what the new world using AI’s potential will look like. This vision and strategy will then support the realisation of the roadmap.
Align use cases with ‘value pools’
Finding use cases aligned to “value pools” is important, Candy said. Whether it is a bank, a utilities provider or a retailer, Candy suggests asking where in the organisation there are heavy manual knowledge bases or document-intensive activities slowing down cycle times.
“One example is a contact centre, where we are seeing a lot of people focusing on generative AI,” said Candy. “How can you make your agents more effective by wrapping knowledge bases around them to improve call handling times or agent actions to deliver a better experience for the customer?”
Start an innovation engine for AI
Scaling AI requires an innovation engine for organisations to take use cases right through to validation, testing, piloting with minimum viable products and scaling. Candy said the likes of design-led and product-led ways of working could contribute to organisational success.
“You need that agile flywheel to build, deliver and scale,” Candy said.
Build a generative AI tech core
Candy said organisations need to be clear on the architecture and digital core layer they will use to manage their generative AI. This clarity is required because many organisations will be using AI across multiple clouds, in addition to AI-infused products like Salesforce and SAP.
Get on top of AI governance
Enterprises will need to manage problems like bias, drift and explainability across multiple AI models, as well as have the right processes in place for their people. They will also need to be in compliance with regulations being created in multiple legal jurisdictions around the world.
Prioritise the people challenge
About 70% of the challenge of rolling out AI is a human challenge, Candy argues. This includes infusing employees with the skills required to have confidence with AI models and maximising change management success by getting AI adopted widely by employees at the coalface.
The future of AI in Australia will balance use case regulation with innovation
Matthew Candy is “very excited” about the potential of generative AI. IBM Consulting, too, is rolling out generative AI; it recently announced it would augment its 160,000 consultants with a variety of AI assistants to scale expertise across a range of roles in the organisation.
Candy said these AI assistants would be able to encapsulate organisational knowledge and handle more of the repeatable parts of roles. These assistants will also be offered as products and services for IBM Consulting clients to help them scale better and faster through 2024.
Australian AI regulation right to focus on high risk use cases first
Australia’s announced regulatory approach for AI, which follows the European Union in focusing future regulation around the risks presented by specific AI use cases, was the “right approach,” Candy said.
“We believe in making sure there is appropriate regulations and controls in place for different types of use cases; it’s really about regulating where the use case lies,” said Candy.
IBM predicts rise of smaller models, innovation and governance
Enterprise clients in Australia and around the world are thinking carefully about the best AI models to deploy. Candy predicts smaller models will be attractive in 2024 due to advantages like less hallucinations and compute requirements, leading to lower costs.
Another trend to expect in 2024 is the emergence of innovative new digital products and services that could shift and transform existing business models, with Candy expecting a lot of big ideas and visions to emerge this year through excellence in generative AI.
With responsible AI becoming critical to organisations, Candy also predicts there will be a continued focus on AI governance foundations, both from a technology standpoint as well as integration into the people and processes of Australian organisations.