Kensho Link
[Workshop] Unified Company Mapping: Leveraging Cross Reference Services and Kensho Link
Listen to our workshop replay to learn about S&P Global's innovative Blueprint for company linking featuring Kensho's advanced AI for seamless entity recognition. Discover how our BECRS dataset and Kensho Link can streamline the mapping and maintenance process with high precision.
Kensho Link uses advanced Machine Learning (ML) to map your messy company data to S&P Global company IDs and global identifiers, reducing time and effort to connect to the S&P Global data universe.
To connect into S&P company data, customers’ databases of companies need to be mapped to S&P Global company identifiers. To make this process fast and accurate, Kensho Link uses a machine learning model that takes inputted company names and then accounts not only for what company data is submitted, but the contents of the entire S&P Global’s company databases. The model then assesses the best mapping and returns a top match (or up to the top five) with a score that represents its quality on a 0-100 scale. Users have the option to have their data returned not only with the top links and scores, but also with basic company foundation data, like legal name, address, URL, and phone number. BECRS subscribers can even have Link return those identifiers as well.
Kensho Link works at scale and with speed. Users can submit files of up to 100 MB of company data at a time (containing millions of entities) and get their results in less than a few hours; or a single submission through our REST API is typically processed in less than a second. The Link model is ready to go out of the box and users need only submit a company name, though the model can be more accurate with additional information such as address, city, state, country, url, or company name aliases.
Kensho Link is accessed in two ways:
Browser-Based File Uploader (“Drag and Drop”): For users who prefer the ease of a browser-based tool, Link can also be accessed through a drag-and-drop interface where users can upload a csv file of their entities of up to 100 MB (as many as several million entities). Processing times vary depending on the number of entities.
REST API: Users can submit up to 100 entities in JSON through the Kensho Link REST API. Results for a single entity are resulted in under one second and 100 entities will normally process in a little under 10 seconds. For larger jobs, users can submit a CSV up to 100 MB through the API. Processing times vary depending on the number of entities.
Service Provider Information
Input: Entity Information with examples
The table below represents a row in a database. Click the examples to populate the table and then find a match with Kensho Link. You can also type data directly into each cell, to try examples of your own.
| Entity Name* | Aliases | Country | Address |
|---|---|---|---|
Key Information
Use Cases
S&P Global currently uses Kensho Link to assimilate new data into the S&P universe.
- Users that require instant mapping of incoming company names to S&P identifiers quickly enrich their data and gain insights will often use the Kensho Link API–or even integrate the API into their automated workflow.
- For users that regularly manage large volumes of data, Kensho Link can accurately process even millions of entities at a time so they can keep their company databases up to date with the most recent S&P data.
- Companies that are trying to augment their CRMs with S&P data, but first need to map their customer names to S&P identifiers. Or in some cases, Kensho Link can help them on their own by assisting in deduplication of their CRM or other company database.
- Even internally within S&P, numerous data teams actually use Kensho Link for various data ingestion jobs.
Benefits
Link began as an internal project, helping S&P Global more quickly integrate new datasets into its universe before it became a commercial product to assist customers with similar data mapping pain points. To date, the solution has linked nearly 500 million entities for internal and external users.
Users can get quick, successful matches with as little as a name. Though that is the minimum requirement, the more information given to the model, the better the results will be.
Kensho Link lets you unleash the power of your data by reducing manual processes using scalable machine learning technologies.