Entity Resolution
Entity resolution is matching records across systems that refer to the same real-world thing — the same customer, company, or account — when the names, spellings, and identifiers don't line up. It's long-standing data-engineering work, and agents with structured, indexed context can now do a usable version of it without a human defining the join first.
Also known as: record linkage, entity matching, fuzzy matching
This used to be tedious, manual work. Someone had to notice that a field in one system and a differently named field in another meant the same thing, or that a company name was spelled one way in the CRM and another way in the support tool, then hand-build the logic linking them. Agents change the shape of that job, because they can recognize the conceptual overlap between fields and entities themselves.
In practice it looks like an agent correlating a customer across separate tools — resolving an organization in Zendesk to the matching calls in Gong — with nobody defining the mapping in advance. It’s one of the clearer arguments for giving agents structured, indexed context instead of pointing them at raw API access. A raw endpoint gives you rate limits and whatever access shape the vendor happened to build; an index the agent can search gives it something to resolve against.
