AI Talk: Turning market volatility into competitive advantage with data & AI

Views on news 

The global economy has entered a new phase. Energy resources, technology, mineral inputs, defence capabilities, financial infrastructure, and data are now regarded as essential components of sovereignty and self-reliance. Everything from financial standards to technology frameworks, corporate governance rules and national security screening regimes are in flux as governments across the world compete for ways to boost domestic growth and capture emerging technologies. As a result, legal foresight – the ability to anticipate legal change and assess the cross-jurisdictional implications of business decisions – has become a crucial differentiator for boards and executive teams seeking confidence in their decision-making. There are seven ways in which new legislative trends can be turned into a business opportunity: investing in AI, exercising effective oversight of cyber resilience, seizing investment opportunities in defence, restructuring supply chains to address global fragmentation, achieving antitrust approvals for necessary consolidation, managing workforce transformation amid demographic shifts and horizon-scanning for emerging litigation risks. The article suggest that the opportunity cost exceeds the cost of moving on with imperfect solutions and that building better decision-making structures pays off very rapidly.  

 

A new model of demand forecasting 

New ways of demand forecasting leveraging AI and first-party data can improve accuracy by up to 50 per cent. As scenario planning is getting more widely used, prediction is moving from figures to options that businesses can decide to respond to. Accurate and complete data sets provide the foundation for any AI, and self-healing data solutions can help with complementing and cleaning available data. Companies fall into three groups: some of them are data ready, some of them are already ready for agents too, but readiness is not just about data and technology but is also organisational, as users must trust AI-driven technologies to make decisions with their help on decisions impacting P&L.  

 

To make accurate predictions, experts need a 360 degree view of customers – ideally from first party data. Near real-time data plays a key role. But forecasting is a means not an end. Businesses must take into consideration the type of decision making they use the data for, whether it’s customer behaviour related or more strategic, such as the length of contracts with suppliers. To find out whether available customer data is accurate and trustworthy, you must try and run an algorithm with that data. If output is not good enough, see how you can improve it. All this must be done in the context of the business outcome you want to achieve, and you need to work backwards from here to see how the model or the data sets need to be improved.  

 

Agentic AI will mean that the purchasing decisions won’t be made by humans but machines. Agentic AI optimisation is aiming for ranking in the top five or less options that the agentic system considers. Broadly speaking, businesses must optimise their site for buyer intentions rather than products, which offers great opportunities for personalisation. Currently, there are two different approaches. One puts emphasis on AI agents and advanced algorithms, while the other focuses on supply chain design. These are two opposing approaches, and we’ll see in the next few years how it’s going to play out. Another challenge will be to detect when an agent was about to execute a purchase but couldn’t find the product in stock. Offering alternatives or ETA for the product should be therefore incorporated into the purchasing process.

 

Agents will need signals from a lot of data sources such as POS, inventory, weather, which must be delivered to them on a single platform.

 

To achieve transparency of the agents and the type of data each has access to, a governance layer will be an essential component of these systems too.  To get buy-in, AI first should be used to enhance the human and its capabilities. AI’s assistance can be very convincing in inventory planning and tracking, where an algorithm can deal much better with huge amounts of data than a human.  

 

The panel’s advice 

• Demand forecasting and supply chain design are the strongest AI use cases in supply chain management.  
• In most use cases, real-time data is not necessary for analysing and predicting customer demand – near real-time (daily) is fast enough to capture changing trends.   
• Businesses must take three steps to become visible for AI agents: focus on the semantics of product descriptions, ensure the product is ready for a transaction (has access of inventory data etc) and design a purchasing journey for the agent on the brand’s platform.  
• Freshness of data will have priority over completeness of data in agentic commerce.