Landbase
Toolkit · Match
← All tools
T4 · Tool

Match

Identifies the entity in the Landbase graph that corresponds to any external record — CRM, CSV, partner list, web-crawl output — with a confidence score. Returns a resolved entity, a candidate set, or no_match.

Examples

How customers use it

01
Matching a CRM export to the Landbase graph
Drop in a 12K-row CRM export. Match identifies the corresponding Landbase entity for every row with a confidence score, and the match-rate itself flags which rows are dirty source data worth cleaning.
02
Pairing first-party signals with Landbase data
Pair internal intent signals — page visits, product usage, support tickets — with Landbase's attribute pack. Match resolves both sides to the same entity so the join is exact, not a fuzzy guess.
03
Anchoring agentic web research to the graph
Run a web-research agent that returns 800 candidate company names. Match anchors each one to the Landbase graph and surfaces which are already in the CRM, which are net-new, and which fail to resolve at all.
04
Deduplicating across messy source lists
Combine three prospect lists from three sources where the same companies appear under different names — "Acme Cloud Inc," "Acme Cloud," "Acme Co." Match resolves all of them to the canonical entity so the unified list has one row per real company.
Pairs with
T3Name to Domain T5Enrich T2Seed-Expansion T6AI Qualification
Found in
CRM Refresh TAM Mapping
Under the hood+
How it's used

The critical interstitial step in every workflow that uses CRM data, pairs first-party with third-party data, or layers Landbase data with the output of agentic tools outside the Landbase ecosystem. The match-rate itself is also diagnostic — a 30% no_match means the source data is dirty or the universe doesn't overlap.

Why it matters

Without Match, you have two disconnected datasets. With Match, you have a unified record set ready to enrich, score, qualify, and act on.

How it works
  • Multi-signal resolution — domain, normalized name, geography, size class, technographic fingerprint.
  • Candidate generation — followed by confidence-based ranking and resolution.
  • Match-rate reporting — dataset-level summary, surfaced as a diagnostic.
  • Human-in-the-loop — optional pass for ambiguous high-stakes cases.