Personalize customer conversations using Real-Time Customer Profile with unstructured data

In Adobe Experience Platform, you can consolidate your customer data into a unified view by combining data from multiple channels with Real-Time Customer Profile. With support for unstructured data, Experience Platform extends these capabilities, semantically letting you use conversational and other unstructured customer data as part of the profile building blocks.

Understanding Profile with unstructured data understanding

Real-Time Customer Profile currently combines structured data from multiple channels, including online, offline, CRM, and third party, consolidating this data into a unified view that offers an actionable and timestamped account of every customer interaction. However, this doesn’t cover unstructured data ingestion, meaning data from conversations, transcripts, or other multi-modal data (such as PDFs or images) are unavailable to be used with Profile.

Profile with unstructured data lets you have access to unprocessed data, which lets you better personalize conversations since the customer context will continue to be shared across multiple platforms, including Brand Concierge.

Profile with unstructured data provides the following capabilities to Experience Platform customers:

Use cases use-cases

Profile with unstructured data lets you use conversational data as an activation signal for further engagements, which allows you to more easily drive inbound personalizations and outbound activations. As a result, you can use these profiles in a variety of use cases:

Sample flow sample-flow

Consider a company that offers financial services to clients.

In this scenario, one of your clients asks an agent about equities. Since this client is logged in, your agent knows about this user by reading their Real-Time Customer Profile.

This client then continues the conversation by stating “I’ve also been meaning to sit down and really review my retirement savings, just haven’t gotten around to scheduling it.”. By using Profile with unstructured data, Brand Concierge and Experience Platform can extract intent from these conversations, which lets you send personalized responses through the agent.

As a result, Experience Platform can make this client enter the “Financial Planning Interest” audience, and immediately personalize the agent to reply with “Let’s make sure your retirement plan is on track — book a free check-in with Alex Agent.”

Additionally, since this context is shared between all Experience Platform applications, you can continue the retirement conversation in multiple forms. For example, Adobe Journey Optimizer can automatically send the client an email with a retirement check-in invitation with Alex Agent.

Beta prerequisites prerequisites

To use Profile with unstructured data, you must meet the following prerequisites:

If you fulfill all of these prerequisites, contact Adobe Customer Care for more information on requesting enablement for Profile with unstructured data.

Requirements requirements

Before continuing Profile with unstructured data set up, you’ll need to complete the following steps:

Once Profile with unstructured data has been enabled, you can start using this feature by setting up the following steps:

Ingest unstructured data ingest

In order to use unstructured data as part of your personalization process, you’ll first need to ingest that data into Experience Platform.

The method to ingest unstructured data differs, depending on the source of the unstructured data.

curl -X PATCH https://platform.adobe.io/data/foundation/catalog/v2/dataSets/67b3077efa10d92ab7a71858
  -H 'Authorization: Bearer {ACCESS_TOKEN}' \
  -H 'x-api-key: {API_KEY}' \
  -H 'x-gw-ims-org-id: {ORG_ID}' \
  -H 'x-sandbox-name: {SANDBOX_NAME}' \
  -H 'Content-Type: application/json' \
  -d '{
    "extensions": {
        "adobe_unifiedProfile": {
            "semanticProfile": {
                "enablement": {
                    "status": "READY"
                }
            }
        },
        "adobe_dataContent": {
            "channel": "call_transcripts"
        }
    }
  }'

[
    "@/dataSets/67b3077efa10d92ab7a71858"
]

curl -X PATCH https://platform.adobe.io/data/foundation/catalog/v2/dataSets/67b3077efa10d92ab7a71858
  -H 'Authorization: Bearer {ACCESS_TOKEN}' \
  -H 'x-api-key: {API_KEY}' \
  -H 'x-gw-ims-org-id: {ORG_ID}' \
  -H 'x-sandbox-name: {SANDBOX_NAME}' \
  -H 'Content-Type: application/json' \
  -d '{
    "extensions": {
        "adobe_unifiedProfile": {
            "semanticProfile": {
                "enablement": {
                    "status": "READY"
                }
            }
        }
    }
  }'

[
    "@/dataSets/67b3077efa10d92ab7a71858"
]

Configure the profile context

After ingesting unstructured data into Experience Platform, you can configure the profile context to help personalize your conversations. You can configure this profile narrative either using Coworker or using the Experience Platform APIs.

GET /semanticprofile/config/category

curl -X GET https://platform.adobe.io.data/core/semanticprofile/config/category \
 -H 'Authorization: Bearer {ACCESS_TOKEN}' \
 -H 'x-gw-ims-org-id: {IMS_ORG}' \
 -H 'x-api-key: {API_KEY}' \
 -H 'x-sandbox-name: {SANDBOX_NAME}'

{
  "requestId": "b2c3d4e5-f6a7-8901-bcde-f23456789012",
  "result": [
    {
      "categories": [
        {
          "name": "Customer Demographics",
          "attributes": [
            "person.name.firstName",
            "person.name.lastName",
            "person.gender",
            "person.maritalStatus"
          ]
        },
        {
          "name": "Contact Information",
          "attributes": [
            "personalEmail.address",
            "mobilePhone.number"
          ]
        }
      ]
    }
  ]
}

POST /semanticprofile/config/configuration

curl -X POST https://platform.adobe.io/data/core/semanticprofile/config/configuration \
 -H 'Authorization: Bearer {ACCESS_TOKEN}' \
 -H 'Content-Type: application/json' \
 -H 'x-gw-ims-org-id: {IMS_ORG}' \
 -H 'x-api-key: {API_KEY}' \
 -H 'x-sandbox-name: {SANDBOX_NAME}' \
 -d '{
    "data": {
        "name": "CustomerSummary":,
        "configurationDetails": {
            "category": [
                {
                    "name": "Customer Demographics",
                    "attributes": [
                        "person.name.firstName",
                        "person.name.lastName"
                    ]
                }
            ],
            "segments": {
                "limit": 1,
                "selection_mode": "inclusion",
                "inclusion": [
                    {
                        "id": "b2cf20ca-5d07-4ef2-b50a-994eb2c11349",
                        "priority": "1"
                    }
                ],
                "exclusion": []
            },
            "marketingActions": [],
            "narrative_precision": "fuzzy",
            "personNamespace": {
                "isPersonNamespaceEnabled": true,
                "code": ["Email"]
            }
        }
    }
 }'

{
    "requestId": "aaaa479b53a61598",
    "result": {
        [
            "CustomerSummary"
        ],
        "configurations": [
            {
                "name": "",
                "configurationDetails": {
                    "category": [
                        {
                            "name": "Customer Demographics",
                            "llm_desc": "",
                            "attributes": [
                                {
                                    "field": "person.name.firstName",
                                    "xdm_desc": "First Name",
                                    "llm_desc": ""
                                },
                                {
                                    "field": "person.name.lastName",
                                    "xdm_desc": "Last Name",
                                    "llm_desc": ""
                                }
                            ]
                        }
                    ]
                },
                "segments": {
                    "limit": 1,
                    "selection_mode": "inclusion",
                    "inclusion": [
                        {
                            "id": "b2cf20ca-5d07-4ef2-b50a-994eb2c11349",
                            "priority": "1",
                            "name": "High Value Customers",
                            "description": "Customers with LTV > $500",
                            "totalProfiles": 125000
                        }
                    ],
                    "exclusion": []
                },
                "marketingActions": [],
                "narrative_precision": "fuzzy",
                "personNamespace": {
                    "isPersonNamespaceEnabled": true,
                    "code": [
                        {
                            "code": "Email",
                            "name": "Email"
                        }
                    ]
                }
            }
        ]
    }
}

Now that the profile context has been created, you can save the configuration against Brand Concierge or any other customer-facing agent.

Profile context in Brand Concierge profile-context-brand-concierge

In Brand Concierge, select the concierge you want to apply the profile configuration to.

Once you’ve selected the concierge, go the Personalization tab and select the previously created configuration.

Profile context with other customer-facing agents profile-context-other

To use profile context with other customer-facing agents, you’ll need to use the Experience Platform API to look up the customer’s profile context and to retrieve the conversation history. By retrieving both of these values, your agent can properly retrieve the full and relevant context before starting a conversation.

To look up the customer’s profile context, you’ll need to make a GET request to the /semanticprofile/lookup/narrative endpoint while providing the entityIdNS, entityId, and configurationName as query parameters.

curl -X GET "https://platform.adobe.io/data/core/semanticprofile/lookup/narrative?entityIdNS=Email&entityId=jane.doe%2B1@example.com&configurationName=CustomerSummary" \
  -H "Authorization: Bearer {ACCESS_TOKEN}" \
  -H "x-api-key: {API_KEY}" \
  -H "x-gw-ims-org-id: {ORG_ID}" \
  -H "x-sandbox-name: {SANDBOX_NAME}"

{
    "requestId": "7f3a9c8e-4d7b-4a6e-9f2c-1b6e8d4a2c91",
    "narrative": "Jane Doe is a member of the 'High Value Loyalty' audience. Her primary email address is jane.doe+1@example.com."
}

To look up a customer’s previous conversation history, you’ll need to make a POST request to the /semanticprofile/query/v1/profile endpoint, while providing the entityIdNS and entityId as query parameters.

curl -X POST "https://platform.adobe.io.data/core/semanticprofile/query/v1/profile?entityIdNS=Email&entityId=jane.doe%2B1@example.com" \
 -H 'Authorization: Bearer {ACCESS_TOKEN}' \
 -H 'Content-Type: application/json' \
 -H 'x-gw-ims-org-id: {IMS_ORG}' \
 -H 'x-api-key: {API_KEY}' \
 -H 'x-sandbox-name: {SANDBOX_NAME}' \
 -d '{
    "conversations": {}
 }'

{
    "conversations": {
        "items": [{
            "conversation_id": "conv-001",
            "channel": "web-chat",
            "created_at": "2026-05-02T14:03:11Z",
            "last_active_at": "2026-05-02T14:21:55Z",
            "participants": [{
                "identity_map": {
                    "Email": [{
                        "id": "jane.doe+1@example.com",
                        "primary": true
                    }]
                },
                "role": "customer",
                "is_author": true
            }],
            "interactions": [{
                "interaction_id": "int-001",
                "timestamp": "2026-05-02T14:03:11Z",
                "contents": [{
                    "content_value": "Hi, I'd like to upgrade my plan.",
                    "content_type": "TEXT"
                }]
            }]
        }],
    "has_more": true,
    "page_token": "eyJjdXJzb3IiOi…"
  }
}

After getting both the narrative and conversations, you can feed this information to your third-party AI agents. As a result, the AI agent will have information about who the customer is as well as information about what the customer discussed.

Configure signals to extract data configure-signals

Once you configure the profile context, you can now configure signals to extract data from your unstructured data.

You can begin this process by going to Coworker and asking “Let me extract signals from conversations”. Coworker then asks what type of signals you want to detect from the conversations.

Next, Coworker asks what type of brand or business these conversations are part of.

Finally, Coworker asks you if you want to create a new signal group or if you want to add to an existing signal group.

You can then confirm if the above steps look accurate, and if you want to proceed with the signal creation.

The signal creation begins. Coworker then creates a signal that automatically gets added to Experience Platform.

Frequently asked questions faq

The following section lists frequently asked questions related to profiles with unstructured data.

Are profiles with unstructured data separate from a normal customer profile?

What is the profile context/profile narrative?

What is a signal?

How does a signal differ from the raw conversation?

Does Profile with unstructured data only support Brand Concierge?

What data sources can I use?