Software Technologies Inventory and Attributes

Section A — Sourcing and Harvesting

Before building the Software Technologies Inventory from scratch, assess whether all or part of it can be harvested from systems already operating in the enterprise. Software technologies are unusually well represented in existing systems, because deploying them leaves traces. Common sources include configuration management databases and IT service management platforms, which often already hold technology configuration items; package managers and dependency manifests, which enumerate what applications actually pull in; container registries and image scanners, which reveal what is running in containerized environments; and software discovery or asset management tooling, which detects installed software across the estate. Harvesting from these reduces the initial data entry burden, accelerates time to Crawl completeness, and surfaces technologies that manual discovery would miss — particularly the ones no one remembers deploying.

AI agents can also serve as an effective harvesting tool — particularly for Software Technologies whose defining information is publicly available or widely documented. For inventories grounded in reference data — regulatory bodies, market sectors, industry-standard capabilities, publicly listed vendors, geographic jurisdictions, and similar Noun Types — an AI agent can be prompted to generate an initial set of Noun Instance records drawn from publicly available sources, producing a draft inventory that practitioners validate and extend rather than building from scratch. This approach is especially effective at Crawl maturity, where the goal is breadth and completeness over depth. Practitioners should treat AI-generated records as a starting point requiring human validation — not as authoritative records — and should document the generation method in the Provenance and Audit Attributes category of each AI-generated record so its origin is transparent and auditable.

Where harvesting is not possible — because no source system tracks this Noun Type, because source data quality is insufficient, or because the attributes required go beyond what any source system captures — the inventory must be built and maintained manually. Manual inventory governance is entirely viable at Crawl and Walk maturity. It requires discipline, clear ownership, and a regular reconciliation cadence — but it is not a failure mode. The most important thing is that the inventory exists, is accurate, and is actively maintained. The tooling and automation can come later.