Lack of trust is holding back the adoption of AI in supply chains

Despite all the hype surrounding artificial intelligence, the actual adoption of AI in supply chains is still proceeding at a strikingly slow pace. Only 2 per cent of companies have AI applications that operate fully autonomously. A lack of trust is the biggest barrier, followed by insufficient data quality and a lack of clarity regarding responsibilities. This is according to international research by the analyst firm IDC, commissioned by supply chain software provider Kinaxis, involving 2,176 supply chain leaders.

AI is now being used on a large scale in supply chain processes, but the move towards autonomous decision-making is proving difficult. According to the report ‘Making supply chain AI accountable’, 56 per cent of companies use artificial intelligence for co-pilot-style applications. A further 54 per cent use it for predictive analytics. For 50 per cent, AI supports automated replenishment and planning, whilst 46 per cent use it to detect and filter out exceptions.

Fully autonomous decision-making, however, is rare. Only 2 per cent of respondents state that their organisation has fully autonomous AI capabilities. At the same time, 40 per cent do expect the supply chain to operate largely autonomously within two years. According to IDC, this gap between ambition and reality is a major challenge for supply chain directors.

Trust is the main barrier

The biggest obstacle to further adoption appears to be not the technology itself, but trust in AI-driven decisions. Of the respondents, 52 per cent cite trust as the main barrier to faster adoption. Users want to understand why AI recommends a particular decision, what data was used to reach that decision, and when human intervention is necessary.

This explains why supply chain organisations are reluctant to let AI make decisions independently. Managers want to retain control, particularly when decisions affect delivery reliability, stock levels, production capacity or customers. IDC emphasises that trust is not built solely through better algorithms, but also through transparency, clear lines of responsibility and demonstrable results.

The ability to demonstrate the value of AI investments plays a key role in this. As many as 62 per cent of respondents say that better data quality and system integration would accelerate investment in artificial intelligence. In addition, 51 per cent cite clearer KPIs for measuring the value of AI. For 50 per cent, simpler integration with existing platforms would be a major incentive.

Reliable data as a prerequisite

Poor data quality is a second major obstacle. Of the respondents, 42 per cent cite problems with data quality and system integration as a barrier to further AI adoption. A further 31 per cent are concerned about the reliability and integrity of data.

According to IDC, AI applications can only make reliable decisions when the underlying data is up to date, consistent and accessible. In supply chains, this is particularly complex because information is spread across ERP systems, planning software, transport management systems and external supply chain partners.

Improving data quality therefore requires more than simply cleaning up databases. Organisations must identify which data is crucial for decision-making, where it comes from, who is responsible for its quality and how different systems are interconnected. IDC regards reliable data as a prerequisite for further automation, not as a problem that AI can solve on its own.

Governance as part of day-to-day operations

Internal oversight – or governance – of AI is also lagging behind technological ambitions. 67 per cent of respondents expect AI governance to change significantly over the next two years. Accountability, risk management and compliance are cited in particular as priorities.

This means that companies must clearly define who is responsible for AI-driven decisions and which decisions may be taken autonomously. In addition, there must be escalation procedures for exceptions, and managers must be able to verify why AI makes certain recommendations.

IDC therefore advocates a shift from governance as a set of policy rules to governance as part of day-to-day operations. This includes clear decision-making rights, human oversight at critical moments and measurable results. AI must not only make decisions faster, but also make demonstrably better decisions.

Greatest impact on logistics and stock management

Expectations regarding the future impact of AI are high. Of the respondents, 46 per cent expect artificial intelligence to have the greatest impact on logistics over the next two years. Inventory management and supply planning follow, also at 46 per cent, whilst 45 per cent cite demand planning. Procurement lags behind at 32 per cent.

It is striking that planning functions, in particular, are already relatively advanced in terms of AI applications, but this does not automatically mean they are ready for autonomous decision-making. The researchers expect AI to add value primarily when organisations can respond more quickly to disruptions, better align stocks with demand, and link planning decisions to execution.

Start with concrete operational results

IDC advises supply chain leaders not to start with the ambition of making processes fully autonomous, but with concrete operational results. It is therefore necessary to define in advance what improvements AI should deliver, for example in service levels, forecast accuracy, stock costs, productivity or decision-making speed. Trust can then be built by testing AI recommendations, analysing deviations and providing users with insight into the underlying decisions.

In addition, according to the analyst firm, companies must invest in data quality and integration, establish clear responsibilities and determine when human approval remains necessary. Autonomy can then be expanded step by step, once reliability and added value have been demonstrated.

 

The post Lack of trust is holding back the adoption of AI in supply chains appeared first on Supply Chain Movement.