The single view of the customer Client Associates didn't have — the same account, mastered in every acquired brand's system, resolved into one.
24 fragmented records across 8 source masters resolve to 10 golden records — a 58% duplicate rate collapsed on merge. Reconcile the 6 customers with conflicting payment terms to stop cash leakage and unlock the 9 cross-brand accounts.
10 of 24 records resolved to golden · 14 duplicates collapsed · 6 term conflicts open
The same buyer sits on Net 30 in one brand and Net 60 in another across 6 accounts — direct cash leakage and credit risk a single master removes.
24 raw records reduce to 10 golden — finishing survivorship gives every team one account, one team, one set of terms.
9 golden records are already served by ≥2 brands — the resolved master is the cross-sell base no brand system could see alone.
Golden record is the join key for Customer 360 and Cash 360
Every acquisition brought its own customer list, so Sinha & Co., Gupta Enterprises and Kaiser each exist several times across the brand systems — with different names, IDs and even payment terms. Until they're resolved, no one can see the true relationship, total exposure, or cross-sell base. This is the golden record AI builds on top of the mesh.
The ★ master survives; the rest merge with a match confidence. Conflicting terms are flagged in red.
| Source system | Name as stored | ID | Terms | AR | Match |
|---|---|---|---|---|---|
| ★Client Associates CRM (Salesforce) | Sinha & Co. Corporation | ACCT-10231 | Net 30 | $6.1M | master |
| 0AFA legacy (ERP) | Sinha & Co. Corp | AFA-4471 | Net 45 | $4.4M | 96% |
| 0RFI estimating | Federal Express | RFI-2208 | Net 30 | $3.9M | 92% |
| 0Ion247 PSA | Sinha & Co. Ground | ION-882 | Net 45 | $4.2M | 88% |
Conflict: payment terms differ across systems (Net 30 vs Net 45). Survivorship keeps the master's Net 30; the others are flagged for credit to reconcile — exactly the kind of leakage a single view removes.
Cross-brand signal: Sinha & Co. is served by 4 Client Associates brands. That's a relationship to coordinate (one account team, one set of terms) — and a cross-sell base already inside the house.
Fuzzy + LLM matching on name, address and tax ID links 'Federal Express', 'Sinha & Co. Corp' and 'Sinha & Co. Ground' to one entity — with a confidence score, not a blind merge.
Records above the confidence threshold collapse into one; borderline matches are queued for a human steward.
Survivorship rules pick the winning value field-by-field (most-recent, most-complete, or the governed source) to form the golden record.
Conflicts — different terms, duplicate AR — are surfaced to credit and sales, and total exposure is netted across the brands.
The golden record is the join key for everything else: it lets Customer 360 show one account, Cash 360 net one exposure, and the cross-sell engine see that a customer already buys from two brands — the same entity-resolution layer that powers the office-grain and quote views.