Use AI to bridge identity gaps across technology inventory records

Overview

Technology names are not standardized across organizational data sources. The same technology may be referred to as “Python,” “Python 3,” “Python 3.11,” “python3,” “CPython,” and “the Python runtime” in different data sources, by different teams, and in different organizational contexts. When the Technologies Inventory family is analyzed alongside application records, expense records, and other organizational data, these naming variations create identity gaps — cases where the same technology appears under different names in different records and cannot be automatically connected without a resolution step. AI tools are particularly effective at resolving these identity gaps through their ability to recognize that different names refer to the same underlying technology, in the same way that the APM AI-assisted analysis capabilities described in the APM Best Practices document resolve identity gaps in the Applications Inventory.

Best Practice

Apply AI identity resolution as a standard step in the Technologies Inventory population and maintenance process whenever inventory data is sourced from organizational data sources that use different naming conventions than the Technologies Inventory family. Provide the AI with the canonical technology names and semantic identifiers from the Technologies Inventory family and the data from the organizational data source being analyzed, and query the AI to map each technology reference in the source data to the canonical Technologies Inventory record it most likely corresponds to, flagging cases where the mapping is ambiguous for human review. For ongoing currency, run the identity resolution process periodically against updated source data to identify new technology references that may correspond to existing inventory records under different names, or that may represent genuinely new technologies requiring new inventory records.

Benefit(s)

AI identity resolution eliminates the data integration barrier that naming inconsistency creates between organizational data sources and the Technologies Inventory family. Technology adoption data from application records, expense data from finance systems, vulnerability data from security tools, and utilization data from software management platforms can all be connected to the Technologies Inventory without requiring each data source to adopt the inventory’s naming conventions, because the AI resolves the naming differences rather than requiring the source systems to eliminate them.