case study

Improving Supply Chain Risk Insights With Contingent

Hear from Raj Wilkhu, CTO and Co-Founder of Contingent AI, on how his team used Diffbot to expand their picture of supply chain risk.

The biggest benefit of working with Diffbot is a stream of news with resolved entities that match our knowledge graph in a very easy way. The Diffbot KG API is quite easy to use. ... Highly recommend anyone looking to work with Diffbot to get started.

- Raj Wilkhu, CTO and Co-founder, Contingent AI

Contingent AI

About Contingent AI

Contingent AI is a AI startup working at the cutting edge of supply chain risk monitoring and mitigation. Specifically, they are developing the first SaaS solution meant to provide up-to-date supplier (and supplier of supplier) analysis to de-risk exposure, help firms comply with regulation, and prevent fraud.

Contingent AI’s team consists of a team of financial risk technology veterans and AI researchers operating at the bleeding edge of Supply Chain Risk, Data, and Computer Science and Compliance/Due Diligence/Risk Screening. Within global supply chain-related news mentions, minutes and accuracy of insights matters deeply. We’re honored Contingent has chosen Diffbot’s Knowledge Graph to help provide topical risk insights structured from the public web to their growing customer base.

The Problem

Contingent AI’s project centered around the issue of how they could identify companies and individuals to link up additional data sources related to supply chain risk. Led by CTO and Co-Founder Raj Wilkhu, Contingent was tasked with disambiguating key organization and person entities at scale from a variety of internal and external (e.g. Diffbot) data streams. Additionally, while Contingent’s internal supply chain-related data contains a host of data points on risk, they hoped to boost news coverage with resolved entities to spell out a “clearer picture” of supply chain risk.

Issues with common approaches to solving this type of issue include industry-specific data stores that aren’t easily integrated into a knowledge graph (entity and relationship) structure like Contingent AI utilizes. Additionally, running with news mentions straight from the web requires time consuming, error-prone, and expensive time structuring data, establishing proxies, and scraping at scale.

Types of web-derived facts feeding into supply chain risk insights:

The Solution

The billions of organization and person entities within Diffbot’s Knowledge Graph provide unique identifiers for disambiguation and the ability to precisely retrieve a range of fields for enrichment or knowledge graph seeding. In the case of Contingent AI, linked data between article, person, and organization entities was central to project speed and success.

The ability to pair up internal and Diffbot-sourced entities important to supply chain risk opens the “flood gates” for a range of connected fact types including key individuals, news mentions, sentiment of mentions, and a host of firmographic details.

Diffbot’s global news index enabled Contingent to quickly expand risk assessment output for their customers and provide scaffolding for their internal data sources.


Using Diffbot for their data acquisition, structuring, and augmentation from web-based sources allowed Contingent AI to jump right to continued experimentation on their cutting-edge supply chain risk platform. Uniformly structured public web data helped Contingent AI to quickly improve data coverage and quality from a new and existing (but unlinked) data sources. Due to the intuitive structure of Diffbot’s Knowledge Graph API, all of this was possible without changing the structure of Contingent’s internal knowledge graph.