Article: Avoiding implementation pitfalls and maximising ROI with Generative AI

As with any new technology, it can be incredibly difficult to forecast the coming impacts of generative artificial intelligence (Gen AI). Expert opinions range from gloomy predictions of an AI “jobocalypse”, to buoyant optimism, and everything in between. What’s certain is that the global economy is in for a period of adjustment.  

According to Deloitte, more than a quarter of the Australian economy is set to be “rapidly and significantly disrupted” by Gen AI, equal to nearly $600 billion of economic activity. By 2030, the number of daily users is expected to double, while businesses are expected to ramp up their annual investment in AI to seven times the amount invested today.

Ahead of Quest’s hugely popular Generative AI Summit 2024 (13-15 August, Sydney), we asked four experts to share their insights into future applications for Gen AI, its impacts on jobs, implementation pitfalls to avoid, strategies for maximising ROI, and planning ahead for related skills gaps.

1. In your view, what are the most exciting future applications for generative AI?

Dr Artak Amirbekyan, Head of Enterprise Data, AI and ML at Transurban, firmly believes Gen AI has the power to transform most of the industries, creating seamless and natural interactions between humans and machines. For instance, the manufacturing industry by simplifying and automating every stage of production from concept to completion. “With Gen AI, humans can concentrate on and even enhance their creative capabilities while leaving the rest to automation, resulting in significantly reduced costs and increased efficiency”, he says. “By streamlining the entire process, Gen AI enables businesses to focus solely on innovative ideas, allowing them to bring products to market faster and more effectively.”

Evolved.AI CEO Dr Michael Kollo points to Gen AI’s potential to revolutionise customer engagement and communication within businesses. “It serves as an advanced communication engine that seamlessly connects firms with their suppliers, staff, and customers. Recent demonstrations by OpenAI and other services highlight its capability to transform interactions and provide compelling communication layers between a business's products, services, and its current and future customers”, he adds.

“I am really excited to see how Gen AI will catalyse research breakthroughs and help find answers to problems we have not been able to solve so far”, says Dr Nandita Sharma, Director of Data Products and Cloud at the Australian Tax Office. “There is also enormous potential in personalised education approaches and tailored learning journeys; especially for those who are underprivileged and do not have access to high quality education”.

2. Vinod Khosla recently said AI has the potential to replace as much as 80% of tasks in 80% of existing jobs within the next decade – do you agree?

“I think that’s a very generalised view”, says Akanksha Wangnoo, Executive Data Risk Governance at NAB. “My opinion is that we need to ensure AI complements the way we work in any industry and helps improve productivity and efficiency – i.e. making our jobs simpler, better and faster, so we can focus on the value-add or real complex tasks that matter most to our customers.”

Wangnoo adds that every technological advancement has created changes in the way in which employers and employees function: “The way we respond to this is in the use cases we select, and our focus on bringing in a lens of improved customer and colleague experience - along with ensuring we support our people with education and learning for emerging technologies such as Gen AI.”

Kollo strongly disagrees with the notion that AI can replace 80% of tasks in 80% of jobs. “Such statements are based on rudimentary models that oversimplify how tasks interrelate within the workplace. These models fail to consider the collaborative nature of work and the interconnectedness of tasks. While generative AI has significant potential, particularly in white-collar roles and communication, it is misleading and potentially disruptive to sensationalise these figures.”

Similarly, Amirbekyan warns that despite a plethora of predictions, it’s challenging to assign an exact number. “Elon Musk recently suggested that all jobs could be replaced by AI. While I can't agree or disagree with his specific estimate, I do believe that many jobs will indeed be automated in the future. It's only a matter of time before technology takes over tasks that can be streamlined. However, as history has shown us, when one job is automated, it often creates new opportunities for innovation and invention. So, while AI may reshape the job landscape, I'm more optimistic about its overall impact on employment. What I am uncertain about is how society will adapt to these changes and the broader implications of AI on our social fabric.”

“This [80% of tasks in 80% of roles] seems quite ambitious”, notes Sharma, “particularly when you consider that many are still struggling with getting access to digital services, internet and high quality data that are key for AI. However, there is significant potential to automate mundane tasks in digital based jobs if there is more confidence with the uptake of automation and AI with a clear governance and risk management framework.”

3. What are the main pitfalls organisations should avoid when implementing generative AI?

“The key pitfall to avoid is jumping straight into the technology without a clear understanding of the business use case and value”, says Sharma. “In addition, businesses should avoid investing in Gen AI without a clear strategy, governance framework, and risk framework. Ethics principles and guidelines need to be in place as well.”

Wangnoo agrees that failing to prioritise ethics principles is a major pitfall facing organisations implementing Gen AI. “Having the right governance, frameworks and guardrails in place to support AI and Gen AI is critical for ensuring ethical and responsible AI use”, she says. “Clarity regarding AI involvement becomes particularly vital in sensitive contexts.”

Wangnoo adds that organisations should ensure they have a ‘human in the loop’ in scenarios where human intervention is necessary, such as complex and emotionally charged use cases.

Amirbekyan warns against assuming and expecting that every Gen AI use case will bring monetary benefits. “This is a new field, and a lot of things are unknown”, he says. “So, having an experimentational mindset is really important.”

For Kollo, the main pitfall is taking too little risk. “Organisations should embrace the fast-evolving nature of technology and be willing to experiment”, he urges. “Expecting to find a stable, easy-to-implement solution without risk is counterproductive. Innovation requires taking risks and understanding that current implementations will continue to evolve and change over time. Avoiding risk and sticking to utility solutions stifles the potential for significant innovation.”

4. What strategies have you seen for getting the highest possible ROI out of Gen AI?

Kollo’s message is to be bold when thinking about Gen AI and ROI. “Traditionally, ROI calculations focus on time savings, measuring how generative AI can reduce hours spent on tasks like reporting or information gathering”, he says. “While this can yield significant savings, a more exciting opportunity lies in using generative AI to create new business avenues and revenue streams. This requires imagination and risk-taking but can drive growth and unlock new opportunities beyond mere efficiency gains.”

Amirbekyan advises leaders to take the time to identify the potential benefits and value-adds before embarking on a Gen AI project. “This could involve collaborative brainstorming sessions with finance professionals and business owners to understand how the project can drive tangible results and create positive impact”, he says.

“Gen AI return on investment can be maximised with a strategy which has a good combination of short-term optimisation and productivity boosting initiatives, and long-term business-transformation initiatives”, says Sharma. “It should involve continuous iterative approaches with clear business metrics to track and measure success; including focused business use cases and incentives for a ‘fail-fast, fail-forward’ culture.”

5. What do you see as the main AI-related skills gaps that organisations need to hire or train for?

In Sharma’s opinion, AI Business Integration and AI Engineering are the two spaces that need significant uplift. “This will help organisations really bring scientific AI to mainstream business and help businesses to realise its full potential”, she says.

Wangnoo believes organisations should ensure that everyone understands the big picture for Gen AI, its associated ethical/privacy/security implications, and deliver tailored education for senior management on what the strategy is and what it means for the business, so that they can lead the charge. She adds “NAB has been really strong on creating a culture of education and learning. Five years ago we established a data and analytics guild to help train and develop our people with their technical and conceptual skills across data science practices, data governance, data privacy and literacy, and the ethical use of data, AI and Machine Learning.”

Kollo notes that generative AI requires different skills compared to traditional AI, which focuses on data science and engineering: “Key skills for generative AI include conceptual and critical thinking, cognitive frameworks, creativity, and imagination. Organisations need to foster strategic and conceptual skill sets rather than operational ones. This shift will necessitate a new educational culture, training individuals to think conceptually and effectively use generative AI, which is an area of significant interest and focus for future development.”

Amirbekyan also believes we have reached a critical juncture in terms of skillsets. “The need is no longer for specialised AI experts, but rather for upskilling all employees to be proficient in using AI tools and technologies as part of their daily workflow”, he says.

Continue the conversation at the Generative AI Summit 2024.

Join us at the Generative AI Summit 2024 on 13 -15 August at the Aerial UTS Function Centre in Sydney to hear from an incredible lineup of industry leaders and AI trailblazers. View the full agenda here.

To access the detailed conference program, download the brochure here.