Four adjacent categories do parts of this well. None of them do the fusion.

What they do well

Excellent at finding what's written down. If the answer exists in your documents, these tools will surface it faster than any human.

Where they stop

They can only search what exists. The workaround nobody documented, the dependency in someone's head, the process that actually runs, none of it is in the corpus, so none of it is in the answer.

What they do well

Strong at measuring communication patterns at scale, who emails whom, meeting load, network centrality.

Where they stop

Volume is not value. Communication metadata can't tell you what a process is, who keeps it alive, or what breaks without them, and scoring people from metadata is exactly the surveillance employees rightly fear.

What they do well

Very good at detection: spend, licences, external assets, exposed services. The technical inventory problem is largely solved.

Where they stop

Detection without ownership is just anxiety. These tools can't tell you why the tool exists, who depends on it, or whether retiring it breaks payroll.

What they do well

Powerful where event logs exist: it reconstructs the real flow through ERP and ticketing systems with rigour.

Where they stop

Most of the company doesn't happen in logged systems. The spreadsheet, the phone call, the person who just knows, process mining is blind to all of it.