Removing bias from supply chains: Using technology to improve how the world does business

March 15, 2021
Vinay Kapoor

By Vinay Kapoor, Interos VP of Product Management

Imagine for a moment, that you are in an exotic new land on vacation. I know it is hard to do this during a pandemic when we are locked inside our homes, but let’s do our best and let our imagination run wild. Now imagine that you get hungry and are looking for a nice place to grab some lunch. What would you do? Would you send a survey out to all the restaurants in the area so they can report how good their food is? Or would you rely on something smarter? The shocking truth is that while most people would not trust self-reported data full of human bias for their lunch, most supply chain modeling and third-party risk management is done in this manner.

Back in the day, when you needed to find out who you did business with and if your supply chain was resilient, you would have relied on survey data or scoring that was run with Excel files under the hood. Furthermore, the best you could do was to get that information down to one-tier. You may have known who you were directly related to, but sub-tier management was impossible — in other words, you had no information on how deep the rabbit hole of your supply chain was and where it led.

At Interos, we have worked very hard to fundamentally change that approach and to help businesses understand their relationships with one another in a way that represents a departure from the status-quo — a new business model and a new way to solve this age-old problem. We use an automated platform that employs artificial intelligence and sophisticated models to find relationships and risks that even our customers did not know existed. We do it to a level deep enough that allows our customers to truly map and understand all of their business relationships in totality. The reason we went down the path of using AI was because we saw an inefficiency and bias that was introduced by the primarily manual approach of the past.

“If humans and machines are both error-prone, then what can we do to build a better solution? ”

It is well known that as humans, we are prone to a number of cognitive biases. Not only do our conscious and unconscious biases impact our work, but they also tend to make us error-prone. For us, it was clear that our customers could do better than relying on third-party risk management data that were incorrect and biased. We saw technology as an answer to this question.

We also recognized, however, that technology alone cannot be the perfect answer. Algorithms, too, are prone to bias, primarily driven by the quality of data being fed into the system. For example, an undergraduate at Georgia Tech found out that a robot she was trying to work with literally could not see her, because it had been trained with data from photographs of lighter-skinned people.  Where does this lead us? If humans and machines are both error-prone, then what can we do to build a better solution for supply chain modeling?

We ask ourselves these questions every day at Interos. The answer to these questions has led the world’s first AI-driven platform for supply chain resilience that features a human-in-the-loop design, a white-box approach to AI, and a robust methodology team that is dedicated to constantly improving our models for continuous improvement and removal of bias.

The solution to the problem we are trying to solve is not easy. Data on business relationships is not readily available in one or even a handful of data sources. The problem requires combing very large data sets from thousands of data sources that are both structured and unstructured. It is a problem that is uniquely difficult for a human-only or a machine-only approach and our solution elegantly combines both in a proprietary manner, including deep learning and cutting edge AI and Machine Learning models to help customers gain complete situational awareness, enabling comprehensive sub tier management.

The agile Interos platform allows our customers to dig several tiers deep into their supply chain and then identify risks, such as finance, cyber, or operations, based on a proprietary graph and supply chain modeling that literally take the guesswork and excel out of the equation. We allow our customers to make third-party risk management decisions that are driven by solid science and data, with the confidence that these decisions are free from bias and errors that you would see in the old, outdated way. In addition, our customers can even get alerts when a relevant event causes a ripple effect that can hit them in the future, well before it hits them. Our near real-time alerting system constantly keeps an eye on a number of well-defined events around the world and gives our customers an early warning system for optimal sub-tier management.

While all this can sound very exciting, what is even more exciting is that we are just getting started. With the power of our proprietary graph of business relationships in the world, our automated third-party risk management models, and our robust near-real-time alerting based on global events, the possibilities are endless. Just like the joy from a nice meal, at your new favorite spot at that dreamy destination.

To learn more about how you can better protect your supply chains, click here. 

View next

Ensure Operational Resilience

Request Contact

Build operational resiliency into your extended supply chain:

  • 889 compliance – ensure market access
  • Data sharing with 3rd parties and beyond – protect reputation
  • Concentration risk – ensure business continuity
  • Cyber breaches – assess potential exposure
  • Unethical labor – avoid reputational harm
  • On-boarding and monitoring suppliers – save time and money