Customer Profiles is the Amazon Connect capability that combines data from historical interactions within Connect and from third-party systems — CRMs, marketing automation tools, business applications — into a single unified customer profile. [Source: AWS, Deliver personalized customer experience using Amazon Connect Customer Profiles] When a contact comes in, that profile surfaces to the agent automatically, so they're not starting the conversation blind.
Where the Data Comes From
Customer Profiles isn't limited to data already inside Amazon Connect. It's designed to combine data from over 75 sources, including leading Customer Data Platforms (CDPs), engagement platforms, and sector-specific applications, through automated ingestion and unification. [Source: AWS, Your unified view of travelers and guests from Amazon Connect Customer Profiles] It also has prebuilt connectors for common systems and can ingest custom data directly from Amazon S3 without needing a bespoke integration built first. [Source: AWS, Create and ingest customer data into Customer Profiles]
How the Data Actually Gets Mapped
The mechanism underneath Customer Profiles is called object type mapping. You define an object type mapping that links specific fields from a source platform — say, a customer's address and email from your CRM — to fields in the unified profile structure. Customer Profiles uses defined keys to look up an existing profile or create a new one, then populates the standard profile object (and any other targeted objects, such as an order or asset record) from the mapped fields, while also storing the original source data as a profile object in its own right. [Source: AWS, How object type mapping works in Customer Profiles]
AWS has also added generative-AI-assisted data mapping, where Customer Profiles samples your source data and suggests the most appropriate target object type automatically, reducing the manual configuration work of setting mappings up from scratch. [Source: AWS, Generative AI powered data mapping in Connect Customer]
Handling Duplicate and Fragmented Records
A common problem with customer data spread across multiple systems is near-duplicate records — the same person with a slightly different spelling of their name, an old email address, or a changed phone number. Customer Profiles uses machine learning to detect profiles that are likely the same person based on similarity in name, email address, mailing address, and phone number, and consolidates them into a single unified profile rather than leaving the agent to piece together several partial records. [Source: AWS, Amazon Connect Customer Profiles]
What this means for the agent, practically: instead of switching to a CRM tab, searching by name, and hoping it's the right record, the agent workspace already shows the customer's unified history, orders, and prior interactions the instant the contact connects.
Where Customer Profiles Shows Up
- Automatically in the agent workspace, as the "customer information" panel
- As data available to flows, so routing and self-service decisions can use known customer attributes
- As context available to step-by-step guides and agent assist, so recommendations can be personalised to the specific customer
Data Storage Note
When you enable the Data store option for Customer Profiles, it automatically stores ingested data, and this cannot be disabled once turned on — deleting data afterward requires using the DeleteProfile API, or removing the object types or the Customer Profiles domain entirely. [Source: AWS, Enable Customer Profiles for your Connect Customer instance] This is worth factoring into any data retention or privacy planning before enabling it.
Frequently Asked Questions
What is Amazon Connect Customer Profiles?
Customer Profiles is an Amazon Connect capability that combines customer data from contact history and third-party sources — CRMs, marketing automation tools, business applications — into a single unified profile, surfaced automatically to the agent when a contact comes in.
How does Customer Profiles bring data together from different systems?
Through object type mapping: you define how fields from a source system (such as a customer's address or email in a CRM) map to specific fields in the unified profile. Customer Profiles uses defined keys to find an existing profile or create a new one, then populates the profile and related objects from the mapped data.
What happens if the same customer exists in multiple source systems with slightly different data?
Customer Profiles uses machine learning to detect likely-duplicate profiles based on similarity in name, email, mailing address, and phone number, and consolidates them into a single unified profile rather than showing the agent several fragmented records.
Can I bring in data without a pre-built integration?
Yes. You can define data from any source using Amazon S3 and enrich a customer profile without needing a custom-built or pre-built connector, in addition to using Customer Profiles' existing connectors for common platforms.
Conclusion
Customer Profiles is the quiet piece of infrastructure behind "the agent already knows who I am" — pulling CRM, order, and support data from dozens of possible sources into one record, deduplicating it with machine learning, and surfacing it automatically the moment a contact starts. It's one of the clearest examples of how the agent workspace reduces the manual lookup work agents used to do themselves.