Workforce gaps keep supply chain AI stuck in pilot mode

Supply chain technology vendors are offering AI agents capable of benchmarking freight rates and awarding freight based on customer-defined rules. However, according to a recent survey, most supply chain organizations are still confining AI usage to generating insights, with workforce readiness cited as the primary obstacle. The Association for Supply Chain Management (ASCM) and the Digital Supply Chain Institute (DSCI) released a report titled "The People Paradox in Supply Chain AI Integration" on October 14.

The survey revealed that 83% of surveyed supply chain professionals felt their employees lacked the necessary skills to utilize AI effectively. Only 17% of respondents reported a readiness to deploy AI, with these organizations experiencing success rates of 71% for AI projects. The remaining 64% of respondents, primarily in the planning stages or piloting AI integration, reported a significantly lower success rate of 28%.

ASCM CEO Abe Eshkenazi emphasized the importance of addressing workforce skills and highlighted that AI should be integrated with confidence in the technology. AI adoption across 13 core supply chain functions was scored on a scale from 1, indicating no AI use, to 5, representing autonomous execution without human review. The average score was 1.94, with planning functions like demand planning and supply and inventory planning achieving the highest scores of 2.28 and 2.12, respectively.

However, execution-heavy functions like returns and reverse logistics trailed, scoring the lowest at 1.59. Eshkenazi compared AI's current application to a "glorified spreadsheet," suggesting that companies should expand AI's authority to handle routine transactional decisions while maintaining oversight over higher-risk decisions such as customer commitments, sourcing, and major disruptions.

The primary barriers identified in the survey were workforce skills and resources, as well as an unclear business case or return on investment, both cited by 33% of respondents. Poor data quality followed at 25%. The survey did not find executive skepticism among the top hurdles. Eshkenazi compared AI to enterprise resource planning (ERP) systems, stating that the technology solution cannot improve bad processes or bad data.

He emphasized the need for companies to use value stream mapping and data analysis to identify and address problem areas before embarking on an AI project. Headcount reduction should not be the sole measure of AI's return, he noted, and instead suggested focusing on response time – the speed at which a company can identify disruptions, evaluate alternatives, and take action.

Involvement in AI initiatives drops below the C-suite. While 82% of C-suite respondents were leading, implementing, or scaling AI initiatives, only 12% at the director level and 51% of managers and 70% of staff were directly involved. The survey collected responses from 144 individuals, with only 11 being C-suite executives, and the results are presented as directional rather than definitive.

Only 7% of respondents predicted that AI would directly replace workers, while 51% believed it would move people toward higher-value decision-making. Eshkenazi warned that this shift carries a cost, as supply chain professionals traditionally built their judgment through analysis, report building, and spreadsheet work, which now falls to AI.

He emphasized that this combination of powerful technology and unprepared workers is a dangerous one for organizations. Additionally, 92% of surveyed professionals expressed significant concern about cybersecurity exposure through AI-enabled suppliers and trade partners. The average organization had adopted only 2.3 of the six standard cybersecurity practices.

Eshkenazi traced this exposure to the increased visibility following the pandemic, as companies began connecting suppliers, carriers, and brokers on shared platforms, leaving them vulnerable to ransomware and cybersecurity risks, particularly in their tier-three and tier-four vendors where resources are limited.