Spanos: In general, technology and AI or analytics can do two main things for businesses.
Firstly, they can improve the quality of business decision making. AI, using quality data inputs, can make better decisions than manual processes. These can range from deciding where to best employ capital, the best price to sell or buy or solving problems within operational processes. Consistency of decision making can also be improved by using a common methodology. This can help to reduce business risk and improve product quality and customer experience.
Secondly, they can improve business efficiency by automating previously manual tasks. This has been possible for some time with the use of technology alone. More modern AI techniques have expanded the number of business processes to which we can apply such tools. In order to compete, businesses must continually assess where AI tools can add value to their customers, employees and shareholders.
What are some key challenges in AI adoption and how should these challenges be approached?
Spanos: AI is a game-changer, but it’s not without its challenges. Here are some of the biggest hurdles and how businesses can tackle them:
- AI hallucinations and misinformation – AI can sometimes generate insights that seem convincing but are actually inaccurate. That’s why human oversight is key. Businesses need to validate AI-driven insights with expert review and ensure AI models are trained on high-quality, reliable data. Ask ICIS helps solve this by combining AI-driven analysis with expert-verified insights, ensuring businesses get accurate, data-backed answers they can trust.
- Change management and trust – Let’s be honest: AI can feel intimidating, especially when it starts automating tasks people are used to doing manually. The key to a smooth adoption is clear communication and training. Employees need to see AI as a tool that enhances their work. When people see AI as a partner adoption becomes much easier.
- Regulatory and ethical concerns – Transparency is everything. AI-driven decisions, especially in commodity markets, must be explainable and compliant with regulations. Businesses need to ensure AI models are built on clear, structured processes that prioritize trusted data and remove bias. Ask ICIS follows a rigorous methodology, applying the same high standards to every query to provide consistent, transparent, and reliable insights.
By tackling these challenges head-on, companies can fully leverage AI’s potential while maintaining trust, accuracy, and compliance.
Please share some best practices for implementing AI-driven automation in complex industries such as the commodities sector.
Spanos: There are some key principles to bear in mind when implementing new AI-based automations.
The first is trust. The success of any AI program will depend on the trust that users have in the system. This is usually only gained by involving them early on in any system design and testing. Change management is also usually a very important component in any AI program. It is rare that such efforts succeed without an accompanying plan to change the way people complete their own work.
Linked to the first principle, a second is to start small and focus on the big things. Try to break down the problem you are trying to solve into manageable chunks and get prototypes in front of customers early. Sharing raw products can help with explaining how the tools work and to maximize the chances of delivering business value. It also helps to minimize the chances of misunderstandings between developers of tools and the business subject matter experts.
