Identity resolution
Matching rules, keys and conflict resolution across sources.
Data Cloud & data unification
One customer view — without copying the whole organization into the CRM.
Designing a unified data layer: identity resolution, harmonization, access to data from external systems, and feeding processes, reports and Agentforce from a governed source.
Capability map
What's included, and in what order
Data Cloud — Capability Map
Identity resolution
Matching rules, keys and conflict resolution across sources.
Harmonization
Normalizing fields and values so they can actually be compared.
External data access
Viewing data from a remote system instead of unnecessary duplication.
Segments & audiences
Building audiences on a unified base for use in marketing and service.
Process activation
Triggering automation, alerts or tasks based on a data event.
Governance & quality
Ownership, quality metrics and drift tracking.
Background
What actually determines the outcome
Data unification isn't a technical project — it's a series of decisions: what defines a customer entity, which identifiers determine a match, and what to do when sources conflict.
Not every piece of data belongs in the CRM. Some is better viewed from a remote source, and some is better brought in only in summarized form. That decision drives both performance and cost.
Data Cloud becomes especially important ahead of AI: an agent relying on unreconciled data will give answers that sound plausible and aren't correct.
What we do
Areas of work
Harmonization
Normalizing fields and values so they can actually be compared.
External data access
Viewing data from a remote system instead of unnecessary duplication.
Segments & audiences
Building audiences on a unified base for use in marketing and service.
Process activation
Triggering automation, alerts or tasks based on a data event.
Governance & quality
Ownership, quality metrics and drift tracking.
Decision matrix
The decisions that determine the outcome
| Decision | Option A | Option B | What decides it |
|---|---|---|---|
| Copy vs. access | Copy data into the CRM | Access the remote source | Usage frequency and performance sensitivity |
| Match key | Technical identifier | Complex business key | Reliability of identifiers in each source |
| Conflict handling | Most recent source wins | Defined source hierarchy | Whether an agreed source of truth exists |
| Initial scope | Customer entity only | Customer and transactions | Maturity of the sources |
Copy vs. access
- Option A
- Copy data into the CRM
- Option B
- Access the remote source
- What decides it
- Usage frequency and performance sensitivity
Match key
- Option A
- Technical identifier
- Option B
- Complex business key
- What decides it
- Reliability of identifiers in each source
Conflict handling
- Option A
- Most recent source wins
- Option B
- Defined source hierarchy
- What decides it
- Whether an agreed source of truth exists
Initial scope
- Option A
- Customer entity only
- Option B
- Customer and transactions
- What decides it
- Maturity of the sources
How we work
Delivery steps
- 01
Map sources
Which systems hold which entities.
- 02
Define the unified entity
Fields, keys and matching rules.
- 03
Harmonization
Normalize values and formats.
- 04
Match testing
Measure false-match and miss rates.
- 05
Feed consumption
Connect to reports, processes and Agentforce.
- 06
Ongoing governance
Quality metrics and defined ownership.
Keep exploring
Related services and guides
FAQ
Data Cloud — questions we hear often
When is Data Cloud justified?
When customer identity is scattered across several systems and you can't answer basic questions without manual reconciliation. If there's one clear source of truth, a simpler solution will do.
Does it replace a data warehouse?
No. A data warehouse is for broad historical analysis, while a customer data layer is for near-real-time operational activation. In many organizations both coexist and complement each other.
What's the connection to Agentforce?
An agent can only be accurate when the data behind it is unified and governed. Data unification is often a genuine prerequisite for real business use of AI, not an optional enhancement.
How long does identity resolution take?
The timeline depends on the quality of identifiers in the sources. When reliable business identifiers already exist the process is short; when matching relies on name and address, a longer calibration cycle is needed.
Next step
Plan a unified data layer
We'll review your sources and identifier quality, and define a realistic unification path.
Next step
Plan a unified data layer
We'll review your sources and identifier quality, and define a realistic unification path.