Although you can deploy models for language processing, Azure Foundry has existing services related to language such as Text PII Redaction. To use that, go to your Foundry proejct, click Services on the left navigation, then look for Text PII redaction.
Sample input text (from Microsoft Learning):
Tailspin Toys Ltd
Invoice
14 September 1984
Customer:
Margaret Ellis
128 High Street, Reading, Berkshire RG1 2AB
Telephone: 021 685 4215
Item: ZX Spectrum 48K home computer (includes power supply, RF lead, and user manual)
Price: £79.00
Payment received: £79.00
When I click the Detect button, it provides the JSON response below. I also selected Redact PII option that is why there are redacted texts. Notice that the JSON response has redacted text, and entities that are assumed to be sensitive. There is a predictable schema or format when using the service as opposed to deploying your own model and customizing it with custom instructions.
{
"documents": [
{
"redactedText": "*****************\n*******\n*****************\n \n********:\n **************\n *******************************************\n Telephone: ************\n \nItem: ZX Spectrum 48K home computer (includes power supply, RF lead, and user manual)\nPrice: £79.00\nPayment received: £79.00",
"id": "1",
"entities": [
{
"text": "Tailspin Toys Ltd",
"category": "Organization",
"type": "Organization",
"offset": 0,
"length": 17,
"mask": "*****************",
"maskOffset": 0,
"maskLength": 17,
"confidenceScore": 1,
"tags": [
{
"name": "Organization",
"confidenceScore": 1
}
]
},
{
"text": "Invoice",
"category": "Location",
"type": "Location",
"offset": 18,
"length": 7,
"mask": "*******",
"maskOffset": 18,
"maskLength": 7,
"confidenceScore": 0.5,
"tags": [
{
"name": "Location",
"confidenceScore": 0.5
}
]
},
{
"text": "14 September 1984",
"category": "DateTime",
"type": "Date",
"offset": 26,
"length": 17,
"mask": "*****************",
"maskOffset": 26,
"maskLength": 17,
"confidenceScore": 1,
"tags": [
{
"name": "Date",
"confidenceScore": 1
},
{
"name": "DateTime",
"confidenceScore": 1
}
]
},
{
"text": "Customer",
"category": "PersonType",
"type": "PersonType",
"offset": 49,
"length": 8,
"mask": "********",
"maskOffset": 49,
"maskLength": 8,
"confidenceScore": 0.99,
"tags": [
{
"name": "PersonType",
"confidenceScore": 0.99
}
]
},
{
"text": "Margaret Ellis",
"category": "Person",
"type": "Person",
"offset": 61,
"length": 14,
"mask": "**************",
"maskOffset": 61,
"maskLength": 14,
"confidenceScore": 1,
"tags": [
{
"name": "Person",
"confidenceScore": 1
}
]
},
{
"text": "128 High Street, Reading, Berkshire RG1 2AB",
"category": "Address",
"type": "Address",
"offset": 78,
"length": 43,
"mask": "*******************************************",
"maskOffset": 78,
"maskLength": 43,
"confidenceScore": 1,
"tags": [
{
"name": "Address",
"confidenceScore": 1
}
]
},
{
"text": "021 685 4215",
"category": "PhoneNumber",
"type": "PhoneNumber",
"offset": 135,
"length": 12,
"mask": "************",
"maskOffset": 135,
"maskLength": 12,
"confidenceScore": 1,
"tags": [
{
"name": "PhoneNumber",
"confidenceScore": 1
}
]
}
],
"warnings": []
}
],
"errors": [],
"modelVersion": "2026-05-01"
}