Case study·Platform

Data-layer identity resolution

From four systems disagreeing on the same person — to one reviewed entity graph

78%
High-confidence entity matchAcross people, companies, properties and documents; remainder held in the review queue
“The join graph made our messy systems usable.”
Data Owner · Mid-market enterprise
Mid-market enterprise·~1,500 employees; ~280k people, ~22k company and ~9k property records spread across the connected systems·4-week pilot·United Kingdom
Talk to us about this workflowTry Unify
01Pilot envelope
Pilot length4 weeks
First signal7 days
First ROI28 days
Team alongside6 seats · 3 colleagues

4-week data-foundation pilot; matched people, companies, properties and documents conservatively; ambiguous matches routed for human review.

02What it owns
Reports toData Owner, dotted line to the CRM administrator on Salesforce writeback decisions.
Owns
  • Entity match table — a reviewed, append-only join across Salesforce, Outlook, SharePoint, BI exports and the customer system of record
  • Provenance index — every joined person, company, property and document carries a source-system link and the rule that matched it
  • Review queue — ambiguous matches routed for human review with a recommended action and a reason
  • Daily data-quality digest — match rate by entity type, low-confidence joins and any threshold drift
  • Duplicate resolution log — append-only, reviewer-approved record of every merge applied to source systems
Does not do
  • Auto-merging records — only proposes merges; the Data Owner approves before any change reaches a source system
  • Lowering the match threshold — any threshold change is signed off by the Data Owner and recorded
  • Writing back to source systems unattended — every writeback runs through the CRM administrator's approval gate
Done looks like

Anyone in the team can ask 'who is this person, what company, which property, which documents?' and get the same source-linked answer — with ambiguous matches transparently held in the review queue.

03The team
AI teammates3
LenaJoins people, companies, properties and documents across Salesforce, Outlook, SharePoint, BI exports and the customer system of record, only matching when name, contact and reference signals agree.
TheoAttaches a source-system link and the matching rule to every joined entity, so any record on the graph can be opened back to its origin in one click.
MiraRuns the daily data-quality digest, surfaces low-confidence matches and threshold drift to the review queue, and tracks match rates per entity type.
Human team6
  • Data OwnerData
  • CRM administratorRevenue Ops
  • Operations leadOperations
  • 4 Data reviewersData
  • IT engineerIT
  • Unify deployment engineerUnify
04Connected stack
SalesforceOutlookSharePointPower BIExcelCustomer system of record
05What it returned
78%High-confidence entity matchAcross people, companies, properties and documents; remainder held in the review queue
65%Duplicate research time reductionFrom hand-reconciling four systems to one source-linked lookup
120Ambiguous records resolvedReviewer-approved with the Data Owner before merging
  • Day 0
    Co-ordinator sessionStakeholders, entity types, identity-resolution thresholds and the first three answerable questions agreed in one call.
  • Day 7
    First signalSalesforce, Outlook and the customer system of record joined on a strict-match rule; Lena returns the first reviewed people and company match list.
  • Day 14
    Read-only entity graphTheo's source-linked graph goes live read-only for the data and operations teams; SharePoint documents and BI exports added against the match table.
  • Day 21
    Review queue activatedMira's daily digest and the ambiguous-match queue go live; the Data Owner signs off on the first batch of duplicate resolutions.
  • Day 28
    ROI reviewSponsor signs off on match rate, lookup-time reduction and the 120-record cleanup; thresholds raised one notch and a second business unit scheduled.
↗Get started

Want this for your team?

Each design-partner pilot starts the same way: one workflow, the minimum useful context, and a first ROI signal measured in days.

Talk to us about this workflowTry Unify