Master data management · use case

MDM with Knobs

Traditional MDM matches records while EKIP enhances them by treating each attribute as a quantifiable indicator, pinpointing areas of uncertainty.

Selection Knobs Creation Knobs Control Knobs Information Geometry Golden Record Intelligence
Traditional MDM asks
"Which record is correct?"
Rules determine the winner, with the final writer prevailing or the source system's priority taking precedence. You will receive a reconciled record, but without any indication of confidence in its attributes.
vs
EKIP asks
What attributes provide the most accurate information about the actual state of this entity?
Knobs evaluate all attributes based on information density from various sources, ensuring the golden record is not only reconciled but also the most information-rich representation of the entity, complete with confidence scores.
The gap

Same entity. Fundamentally
different treatment.

Both methods begin with three opposing source records. The downstream AI accuracy issues stem from their subsequent actions.

Traditional MDM vs EKIP Knobs comparison ── vs ── Traditional MDM Salesforce Acme Corp (v1) SAP ACME CORP (v2) Billing Acme Corporation Rules engine last-write-wins Golden record Acme Corp · no confidence What you get: A frozen artifact No attribute signal No confidence score Rules disagreed silently Ambiguous entities surface no warning EKIP Knobs Salesforce density 0.91 ↑ SAP density 0.43 Billing density 0.28 ↓ Selection knob scores by signal density Golden record Acme Corp · confidence 0.87 What you get: Confidence score Attribute signal map Living record Resolution queue Ambiguous entities surfaced and ranked
How EKIP acts on master data

Three knob types.
One coherent MDM strategy.

Each MDM challenge - survivorship, enrichment, governance - corresponds directly to a knob type on the existing EKIP platform. No need for new abstractions.

Selection Knobs

Survivorship intelligence

Evaluate each source-system value based on its information density rather than solely considering recency or system priority. The attribute that remains is the one that provides the most relevant outcome signal.

  • Source reliability scoring per attribute type
  • Match confidence thresholds for entity resolution
  • Field-level density ranking across all sources
  • Conflict detection with resolution suggestions
Creation Knobs

Entity enrichment

Find entities with limited attribute information and create specific enrichment plans. Develop datasets to complete missing information accurately, without introducing unnecessary data.

  • Attribute coverage scoring per entity segment
  • Cross-system identity stitching (unified customer ID)
  • Synthetic enrichment for sparse attribute clusters
  • Information density gain tracking post-enrichment
Control Knobs

Governance as policy

PII classification, retention guidelines, stewardship assignment, and regulatory risk - all represented as auditable, adjustable controls instead of individual manual tags.

  • PII tagging at attribute level (GDPR, CCPA, HIPAA)
  • Retention and purge policy enforcement
  • Stewardship assignment and escalation rules
  • Regulatory flag propagation across entity types
Information geometry for MDM

Know exactly where to act
in your entity population.

EKIP analyzes entities based on two dimensions: attribute completeness and source certainty, identifying frontier zones with the most impactful knob action. Explore an entity's attribute profile by clicking on it.

Interactive entity information geometry View a scatter plot showing the distribution of banking entities on the certainty and completeness axes. Click on any point to view its attribute profile. sparse + uncertain deprioritize complete + certain · golden record zone complete · uncertain → enrichment target sparse · certain → coverage target attribute completeness → source certainty → 0 0.5 1.0 0 0.5 1.0

Select any entity to view its attribute profile · Click on the background to exit

Selection knob in action

Attribute-level survivorship
for a banking customer.

Attributes are evaluated separately in CRM, Core Banking, and KYC/AML, with the winning source being the one with the highest information density in that particular area.

Banking customer survivorship knob CRM Core banking KYC / AML Selection knob → survivor Legal name Address Risk rating ID / PAN Rel. manager Priya Sharma density 0.62 Priya K. Sharma density 0.91 ✓ P. K. Sharma density 0.48 Priya K. Sharma core banking (0.91) 12 MG Road, Pune density 0.87 ✓ 12 M.G. Rd, Pune density 0.71 Pune, MH density 0.33 12 MG Road, Pune CRM (0.87) Medium density 0.40 density 0.05 Low-Medium density 0.88 ✓ Low-Medium KYC/AML (0.88) BKRPS1234K density 0.31 BKRPS****K density 0.18 density 0.06 ⚠ conflicted Control knob: PII lock Anil Mehta density 0.83 ✓ A. Mehta density 0.55 density 0.09 Anil Mehta CRM (0.83) highest density → attribute survives Control knob detects low/conflicted sources, flags PII, and blocks auto-merge.
Golden record, redefined

Not a frozen artifact.
A living intelligence profile.

An EKIP golden record is a representation of an entity with the highest information density, including a confidence score, attribute signal map, and prioritized resolution queue.

Traditional golden record
Legal name
Address
Risk rating
ID / PAN
Rel. manager
no confidence signal
no resolution queue
EKIP golden record · Priya K. Sharma
Legal name
0.91
Address
0.87
Risk rating
0.88
ID / PAN
0.31
Rel. manager
0.83
overall confidence 0.76
ID / PAN conflicted · stewardship review required
KYC refresh overdue by 14 days
Closed-loop knob lifecycle

MDM is not a project.
It's a continuous loop.

EKIP's lifecycle consists of six stages that align with MDM, starting with identification of high-conflict entities and culminating in the management of the golden record as a dynamic, trackable asset.

MDM knob lifecycle six stage closed loop MDM knob lifecycle 1 · Discover find high-conflict entities by information gap score 2 · Define specify outcome-relevant attributes per entity type 3 · Instrument connect CRM, core banking, KYC, Snowflake 4 · Prioritize rank entities by information gap in frontier zones 5 · Optimize apply survivorship knobs; enrich sparse entities 6 · Govern PII, retention, stewardship as Control Knobs Discover / Define / Govern Instrument / Prioritize / Optimize
AI context layer

AI agents require more than just a golden record; they also need to have a comprehensive understanding. how much to trust each attribute.

The metadata management view introduces the latest feature known as the 'AI Context Layer', provided by EKIP. This layer offers knob intelligence instead of catalog metadata, with each attribute containing its density score, source lineage, and confidence signal. When an AI agent inquires about a customer, it receives both the value and the epistemic context associated with it.

"What is Priya Sharma's risk rating?"
valueLow-Medium
sourceKYC/AML system
density0.88 high confidence
freshness14 days since last refresh
conflictsCRM shows Medium (0.40)
actionKYC refresh recommended

Start treating master data
as infrastructure.

EKIP enhances entity population with knob intelligence, ensuring optimal golden records instead of simple reconciliation.

Explore EKIP → See the platform