AI In Action: Sector Insights and Implications

Sarah Carney, Microsoft’s Chief Technology Officer, joined Morgan Stanley’s Head of European Thematics, Edward Stanley, to discuss their insights into the Artificial Intelligence (AI) sector, implications of AI usage, and current adoption use cases. While many organisations are and will be looking at ways to utilise AI to help them operate more effectively (soft ROI), interesting use cases have emerged which include sectors using AI to improve their core business. 

There has been a trend towards the increased use of ‘shadow IT’ as employees are secretively using AI applications to help complete work tasks, especially when they’ve been prohibited to do so or otherwise blocked by their firm. Sarah claims that she’s “never seen as much shadow IT as she had in the moment of generative AI”. 

Microsoft carried out a Work Trends Index surveying 31,000 people globally ¹, which looked at how they were using AI. Interestingly, 84% of workers in Australia were using generative AI at work, with 78% bringing their own devices and/or logins. This should ultimately concern organisations due to privacy implications of porting data into a public ecosystem. It also brings to light the slow speed at which businesses are moving to keep up with employee demand. Therefore, having the ability to offer secure generative AI at work is becoming a competitive advantage in the talent war. There’s also interesting trends and cognitive dissonance appearing at the individual level. 

Employees are worried that if they use AI, or admit to using it, they may write themselves out of a job, which may be why organisations haven’t seen the quantifiable uptake or productivity increases they were initially expecting. Organisations may need to consider being more specific with employees regarding expectations of how the extra time they were going to have on their hands in light of the AI-generated productivity uplift can and will be spent.

Microsoft was early to market with their Generative AI solutions with their launch of Copilot. Copilot is capturing the market’s attention. It’s been forecasted that enterprise adoption may take some time due to budget constraints and security concerns. Understanding good data governance within an organisation will be paramount, as well as training workers how to use it. On the other hand, the productivity gains and freeing employees up for deeper work are potentially powerful beneficial outcomes from effective adoption. 

Sarah indicates that organisations should heavily focus on data, data governance and data security. Second to data, would be the ability to set up responsible AI and risk frameworks within the organisation. Organisations need to think about how they are going to test these systems and ultimately be comfortable with them. It is imperative that organisations can fully understand how the unstructured data is organised within their AI ecosystems to ensure there is good and safe use of that data. 

Sarah suggests that she sees Generative AI use cases falling within four buckets, with three of the four involving the organisation and people:

1. How can I make my employees more productive? 
2. Can I make my customers happier?
3. Can I improve operations?
4. How do you bend the innovation curve?

Often, the use cases tend not to revolve around the core business of an organisation but rather a peripheral objective, such as improving the efficiency of a call centre. For most organisations, a call centre isn’t their core business but rather a means of supporting their customers. 

Organisations that have seen really interesting use cases are those using AI to change or optimise their core business. The mining industry is a great example, as they’re using AI to help extract more ore or to predict a particular radius of a mine blast. Every tiny piece of ore can make a material and commercial difference to that mining organisation, so they’re already reaping the benefits of AI. 

Sarah indicates that very few organisations are sitting at the revenue side at the moment, there are a few such as construction, mining and gaming but a majority of organisations have started out with cost optimisation because it’s safe. When it comes to return on investment, most are looking at productivity and softer metrics, instead of harder revenue metrics at this time.  

Singapore seems to be doing well in terms of democratising AI for their people and have really thought about how they will put interesting and unusual bodies in place to help the government help their people scale AI. India has also taken an innovative approach and have seen a huge uptake with a significant number of use cases popping up across their economy. Conversely, in tighter regulatory regimes innovation appears to be restricted, this can be seen in the European markets. Australia tends to take a risk-based approach and success will depend on how Australia applies that risk lens to generative AI moving forward.