Using AI to Elevate Supply Chain Due Diligence? Don’t Forget to Pair It With Human Analysis.

Editor’s note: Samuel Logan, author of this article, is co-founder and CEO of Evidencity, a supply chain research software provider.

As markets expand, regulations evolve and risks diversify, thoroughly assessing potential partners, investments or ventures remains paramount. Due diligence acts as a safeguard, providing a comprehensive evaluation framework to mitigate risks, ensure compliance and uncover crucial insights before critical business decisions are made. It protects against unforeseen challenges and optimizes opportunities in an increasingly complex and disjointed marketplace. But it’s just a start when applied to supply chain transparency.

Identifying the origin of products quickly becomes intricate within multinational supply chains. Most consumers are oblivious to the journeys the products they consume make before reaching their households. From the cotton in jackets to the tantalum in cell phone batteries or the abaca used for tea bags, most raw materials and unfinished parts arrive from the developing world, produced in diverse conditions and local communities. 

In addition, extraction and production sometimes happen under murky circumstances that end-users would not condone if they had all the information. Complete transparency is difficult and at times impossible to guarantee since multiple actors, complex international networks and opaque relationships interlink in the various stages of production.

Understanding the multiple facets of supply chain transparency

In enhancing supply chain transparency, investigative due diligence is indispensable to differentiate between good and bad actors. Imagine the supply chain landscape as a jar full of jellybeans — some are green, some yellow and some red — representing the companies and individuals within supply chains. Due diligence can help spot which beans are green (safe), yellow (concerning) or red (unsafe). But today it’s not enough just to color code. Supply chain transparency requires the additional step of analyzing and understanding relationships between the beans — between individuals and companies — through tackling the data accessibility hurdles businesses face when dealing with vast multinationals or challenging jurisdictions. 

Identifying and monitoring these relationships has become crucial for businesses as regulatory enforcement frameworks evolve, bringing financial and reputational repercussions for companies that fall behind. International sanctions and labor transparency legislation only increase the risk of doing business with the wrong people. In this context, businesses already identify the “red beans” swiftly to avoid trouble. But as supply chains grow in transparency, the consuming public will add pressure on companies to identify and justify individuals and companies in their relationship network.

As a result, regular data scans, often aided by AI and automation, can only be a beginning, especially in the developing world, where data taxonomy is a challenge to navigate. Data scans help flag potential issues, identifying the so-called “yellow beans,” but lacking the depth to dig deeper and ensure the subject is not, in fact, a “red bean.”

Scans can indicate who is connected to whom and why, but human analysis remains crucial to tell the full story. Where scans scratch the surface, human expertise and local knowledge provide a nuanced analysis of the data collected, adding the national, political and social perspectives and helping create links with other subjects. Humans on the ground can access information that in 2024 remains beyond the reach of digitalized, commercialized datasets.

The hybrid approach: Data paired with human analysis

A hybrid approach — one that combines the right data and well-trained and placed researchers and investigators — helps identify red flags quickly while offering the opportunity for a deep-dive into relationships, which require insights that are sensitive to geographical location, local legislation, culture and other elements that could affect the interpretation of findings and access to information. When necessary, an on-the-ground presence makes fact-checking possible through interviews and source inquiries in the local language.

Furthermore, locally based researchers ensure that all publicly available data is accessible, given that jurisdictions in Asia, Africa and Latin America often require in-country proof of presence or a need for someone to queue in line to retrieve information provided by government services, such as corporate records in Mexico and Israel or litigation records in Brazil or Thailand.

As the political and legal conditions around data availability in different parts of the world shift, due diligence professionals must adapt to the new reality. Despite robust data solutions that usually include AI and automation, the human experience can only partially be replaced in situations of high complexity.

In the coming years, due diligence as a checkbox exercise will be just a starting point for those who want to identify, interpret and effectively address adverse issues linked to opaque or muddled relationships in their supply chains. The hybrid approach ensures that the right dots are connected appropriately and that multinational companies build supply chains that contribute to the improvement, not detriment, of local conditions and communities, their brand perception and bottom line.