Blog — 4 April, 2026
Introducing the Next Evolution of Kensho Link
Advanced AI, broader coverage, and multilingual matching. Kensho Link has been
significantly enhanced to meet the scale and complexity of today’s data challenges.
Messy company data is an expensive problem, with data preparation still consuming valuable analyst time. In Alteryx’s The 2025 State of Data Analysts in the Age of AI report, respondents said they spend an average of six hours each week preparing and cleaning data.[1]
Much of the challenge comes from entity data itself: a single counterparty can appear under different names, identifiers, and formats across platforms, leaving teams with manual reconciliation work that cannot keep pace with the volume and complexity of the data.
Kensho Link was built to solve this. The next evolution of the platform brings significantly expanded capabilities: broader coverage, multilingual matching, smarter name recognition, and deeper intelligence across complex corporate structures.
What’s New?
Advanced AI |
70 million+ Companies |
Multilingual Matching |
The enhanced Kensho Link is built on machine learning, natural language processing and large language models. This multi-model approach evaluates potential matches through consensus, improving both precision and coverage, even when identifiers are incomplete or missing. |
Kensho Link now connects to S&P Global’s full company universe of over 70 million public and private entities, nearly doubling previous coverage (up from 37 million.) Coverage gaps that previously required manual workarounds are substantially reduced. |
Matching now handles accented characters and non-Latin characters including Chinese and Arabic. For the first time, mapping requests can be submitted in non-English languages, a meaningful step for organizations operating across multiple jurisdictions. |
Key Features
Abbreviations, aliases and alternate names
Company names in the real world rarely match their legal registration exactly. Kensho Link recognizes common abbreviations, trading names and historical aliases. For example, searching “S&P” returns the correct entity of ‘S&P Global’. And, “SpaceX” resolves to its legal name, “Space Exploration Technologies Corp.” without manual pre-processing.
Parent-child hierarchy resolution
Corporate structures are often complicated. Kensho Link distinguishes between a global parent, its subsidiaries and rebranded entities, so firms managing exposure at the corporate family level get the right match, not just the closest one. This is particularly valuable when navigating post-merger structures or multi-jurisdiction corporate families.
Richer output, including 25+ identifiers
Matched records return metadata — city, country, company type, industry, URL — to speed up downstream validation. Clients with a Business Entity Cross Reference Service (BECRS) subscription can also retrieve over 25 unique global entity-level identifiers and Ultimate Parent data. This allows clients to build a comprehensive entity master file that connects their internal systems to a continuously maintained external data foundation.
Built to integrate and evolve
Kensho Link is available through both a REST API and a browser-based user interface (UI), giving teams the flexibility to integrate matching into existing workflows or use a drag-and-drop experience for larger file-based jobs. Its modular architecture is designed to absorb new techniques and technologies over time, so the platform evolves as the data landscape does.
How It Works
Upload at minimum an internal identifier and company name, with optional inputs such as city, country, and URL. Kensho Link returns each record mapped to S&P's CIQ Company ID along with a link score, and the metadata needed to validate results and move forward. The output is enriched with attributes including full address information, company type, company status, industry, URL, and more. Kensho Link can handle up to 500,000 records in a few hours.
Match thresholds are configurable to your own risk tolerance. The underlying models are anchored to verified S&P data, reducing the risk of inaccuracies that come with purely statistical approaches.
[1] Alteryx, The 2025 State of Data Analysts in the Age of AI (2025)