Neural Notes: Are Australian execs actually ready for AI?

Welcome back to Neural Notes, a weekly column where I explore some of the most interesting AI news of the week. In this edition: Aussie CEOs are bullish on AI (duh) but a new report reveals they may not be quite ready for it yet.

In news that is surprising to approximately no one, Australian C-suites are placing generative AI at the forefront of their business strategies. According to a new collaborative report from Salesforce and YouGov AI Research, 81% consider it crucial to their success over the next three years. For 38%, it’s their top priority.

Interestingly, 40% of respondents placed the successful implementation squarely on the shoulders of CEOs, with another 29% citing CIOs, CTOs, or other technical leaders as being in charge of the rollout.

However, despite this enthusiasm, a significant proficiency gap remains. While 83% of C-suites report using AI tools regularly, only 51% of these executives consider themselves highly proficient in leveraging these tools effectively.

This proficiency gap extends to their teams, with just 53% of C-suites believing their team is highly proficient in using AI for work tasks. This highlights the need for businesses to not only implement AI but also ensure their workforce is adequately trained to use these tools effectively.

The research also reveals that these challenges are not confined to smaller enterprises. In larger businesses with over 1,000 employees, where 65% of C-suites report having a clear AI strategy, data issues, and proficiency gaps continue to pose significant hurdles.

Although 54% of C-suites in these larger organisations have AI tools running on their desktops all the time, this frequent usage does not necessarily translate into effective AI implementation.

Are you ready for it?

Despite the strong verbal commitment, significant challenges remain, especially when it comes to data management and the actual practicalities of AI implementation. Justin Tauber, general manager of innovation and AI culture at Salesforce, observes that while many C-suites are eager to integrate AI, there’s often a gap between their strategic ambitions and their preparedness to execute.

“There’s an interesting contradiction in the data. We saw 92% of C-suites saying they have a generative AI strategy, but it’s still the CEO’s responsibility,” Tauber said to SmartCompany.

Tauber finds it particularly surprising that the CEO is still seen as the primary figure responsible for AI rollout, even in organisations that claim to have a strategy in place.

“Normally, when you don’t have a strategy yet, it’s the CEO’s responsibility to come up with it. Part of the strategy would be having someone to take care of that, or having enough clarity about it that you can distribute responsibilities across the business,” Tauber said.

This suggests that while executives are keen to adopt AI, some may not fully grasp the complexity of the implementation process, leading to a concentration of responsibility at the top that may not be sustainable.

“My hunch is that when people say they have a gen AI strategy, what they mean is they have a gen AI policy. It’s not quite the same,” Tauber said.

“And it’s understandable because being able to anticipate two or three years down the track of how generative AI is going to transform your organisation is really hard.”

The AI data problem

One of the primary issues with AI implementation in businesses is the cold, hard data. This issue isn’t just about the quantity of data but also its quality and context.

In fact, 90% of C-Suite respondents identified the likes of incomplete or outdated data as significant barriers to AI adoption. And 97% of all respondents agree that having holistic and complete data is critical for building confidence in AI tools.

Tauber points to the data problem as a particular issue for companies trying to build their own large language models (LLMs).

“If you’re trying to build your own LLM, data is an incredible problem. How am I possibly going to get enough data to build a reliable LLM for myself? Even getting to the point of fine-tuning an LLM, or building stuff in-house, is a huge challenge,” Tauber said.

He further emphasised that instead of obsessing over the complexities of AI models, businesses should focus on how AI can be integrated into their operations to drive real value.

“There’s a lot of uncertainty about what the final AI stack will look like, but that doesn’t need to be everyone’s problem,” Tauber said.

“We can actually focus on how we would use AI and how it integrates into what we do, rather than how we should build it, or what’s the right way of building it.”

And it’s a good point. AI, particularly generative AI, has ham-fisted its way into the business world in an overt way that we haven’t really seen before.

He draws a parallel to the mobile revolution, where many businesses rushed to create their own apps without considering whether it truly advanced their business goals.

“Everyone remembers when people said ‘Oh, we need our own app.’ But, why? How does it advance your business to have your own app? So you can be present on someone’s phone? If it’s not going to be used, then you’re just playing with the technology for the sake of playing with the technology,” Tauber said.

AI should be about relationships

Tauber argues that the real value of AI lies not in productivity, but in relationship building.

“Which of the relationships that you would like to work better are constrained by people’s ability to grow large amounts of data,” Tauber asks.

“Identify those, and then work out how AI can make those moments that matter in those relationships work better.”

He elaborated on the importance of AI in facilitating better communication and collaboration within organisations:

“There are all sorts of relationships that need to be maintained, not just relationships with customers… but relationships inside an organisation that need to be maintained, where the cost of maintaining those relationships is actually quite high.”

Tauber pointed out AI can help reduce the cognitive load on employees, allowing them to focus on the more nuanced and human aspects of their work.

“AI can help by summarising what’s going on, by making it easier for everyone’s expertise to contribute, and by helping people to engage more effectively with their work and with each other,” Tauber said.

He also suggested AI can be a powerful tool for fostering more personalised and responsive interactions.

“Imagine AI that can help a sales team quickly understand a customer’s history, or help a manager grasp the key points of a project’s progress. These improvements can make a real difference—not just in terms of productivity, but in the quality of the relationships that drive business success.”

In the end, Tauber believes that AI’s greatest impact will be seen in how it transforms the way people work together.

“The most successful AI implementations will be those that strengthen these relationships—by making communication clearer, decisions faster and collaboration more seamless,” Tauber said.

Other AI news this week

Both Apple and Nvidia are considering investing in OpenAI in its latest funding round. This would bring the company’s valuation to over US$100 billion. As a reminder, OpenAI (which started life as a non-profit) has yet to turn a profit since pivoting to having a capped profit arm.
Speaking of Nvidia, its latest financial results revealed over US$30 billion in revenue over the past three years — a 122% YoY jump. Despite this being the fourth consecutive quarter to see triple-digit revenue growth, its share price took a 6% nosedive. Slow growth and production issues have been pointed at for the dip.
On a sadder note, it’s Scam Awareness Week and there’s been stories of AI deepfakes being used in the likes of romance scams in Australia.
Meanwhile, Google has been sticking to the most important generative AI breakthroughs — developing a model that can simulate the original 1993 Doom. What a time to be alive.
And to ends things strongly, GPT-4o and Claude can’t spell ‘strawberry’.