---
schema_version: "1"
id: b4c667f1c177d08563a13f6fa8a437cc5df8bc51c358563c7f0587c840e3c8f2
title: Einstein Trust Layer Region and Language Support
collection: admin
version: current
locale: en-us
portal_url: https://docs.salesforce.com/admin/current/en-us/ai/264-0-0/generative_ai_trust_lang_region.md
canonical_url: https://help.salesforce.com/s/articleView?id=ai.generative_ai_trust_lang_region.htm&release=264.0.0&type=5
source_site: help.salesforce.com
content_type: documentation
content_hash: f999349c97ec00df727c1042e4dc4473f9f312be5564b889f5af44da8c3d56b1
description: The Einstein Trust Layer detects and masks sensitive data such as personally identifiable information (PII) for multiple regions and languages.
products:
  - Agentforce
  - Cross Cloud
release: "264"
---
# Einstein Trust Layer Region and Language Support

The Einstein Trust Layer detects and masks sensitive data such as personally identifiable information (PII) for multiple regions and languages.

### Required Editions

<table class="slds-table slds-table_bordered slds-max-medium-table_stacked slds-m-bottom_small edition"><colgroup><col></colgroup><tbody><tr class=""><td class="slds-cell-wrap" style="vertical-align:top;">Available in: <strong>Enterprise</strong>, <strong>Performance</strong>, and <strong>Unlimited</strong> Editions with an Einstein for Sales, Einstein for Platform, Einstein for Service, Einstein 1 Service, or Einstein GPT Service add-on. To purchase add-ons, contact your Salesforce account executive.</td></tr></tbody></table>

## Field-Based Data Masking

Field-based data masking supports all Salesforce languages. To learn more, see [Large Language Model Data Masking](https://help.salesforce.com/s/articleView?id=ai.generative_ai_mask_data.htm&release=264.0.0&type=5 "The Einstein Trust Layer identifies and masks sensitive information in the prompts before sending them to the large language model (LLM). The Einstein Trust Layer uses data masking to protect sensitive data such as personally identifiable information (PII) from being exposed to external LLMs.") and [Select What Data To Mask](https://help.salesforce.com/s/articleView?id=ai.generative_ai_mask_select.htm&release=264.0.0&type=5 "Select which data you want to mask and prevent exposure of sensitive data to the large language model (LLM). These settings are applied to your Salesforce org.").

## Pattern-Based Data Masking

Sensitive data is identified using two methods:

- Model-based detection: Machine learning models are trained to identify specific data types such as names of people or companies. For example, the model can scan a block of text to detect a name and categorize it as the name of a person or a company.
- Regex and context-based detection: Detection is dependent on how precisely the text matches the regular expression (regex) pattern, the uniqueness of the pattern, and the proximity of relevant context words to the pattern.

These data types are supported in English, French, German, Italian, Japanese, and Spanish:

<table class="slds-table slds-table_bordered slds-max-medium-table_stacked slds-m-bottom_small ">
        <colgroup>
          <col style="width:29.275362318840582%">
          <col style="width:41.73913043478261%">
          <col style="width:28.985507246376812%">
        </colgroup>
        <thead class="slds-text-title_caps ">
          <tr class="">
            <th scope="col" class="slds-cell-wrap" style="vertical-align:top;" id="d70589e103">Data Type</th>
            <th scope="col" class="slds-cell-wrap" style="vertical-align:top;" id="d70589e106">Pattern</th>
            <th scope="col" class="slds-cell-wrap" style="vertical-align:top;" id="d70589e109">Detection Method</th>
          </tr>
        </thead>
        <tbody>
          <tr class="">
            <td data-label="Data Type" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e103 ">Company Name</td>

            <td data-label="Pattern" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e106 ">The name of a company or organization</td>

            <td data-label="Detection Method" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e109 ">Model- based </td>

          </tr>
          <tr class="">
            <td data-label="Data Type" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e103 ">Credit Card</td>

            <td data-label="Pattern" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e106 ">A valid 15, 16, or 17 digit credit card number with or without spacing. For
              example, 4111 1111 1131 1121</td>

            <td data-label="Detection Method" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e109 "> Regex pattern + relevant context words</td>

          </tr>
          <tr class="">
            <td data-label="Data Type" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e103 ">Email Address</td>

            <td data-label="Pattern" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e106 ">An email address. For example, astro@salesforce.com </td>

            <td data-label="Detection Method" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e109 "> Regex pattern + relevant context words</td>

          </tr>
          <tr class="">
            <td data-label="Data Type" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e103 ">IBAN Code</td>

            <td data-label="Pattern" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e106 ">An International Bank Account Number (IBAN) that starts with a two letter
              country code, two numbers, and an account number. For example,
              GB29NWBK60161331926819</td>

            <td data-label="Detection Method" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e109 "> Regex pattern + relevant context words</td>

          </tr>
          <tr class="">
            <td data-label="Data Type" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e103 ">Name</td>

            <td data-label="Pattern" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e106 ">A person’s first, middle, or last name, or a combination. Titles, such as Mr.,
              Ms., and Mrs. are also considered.</td>

            <td data-label="Detection Method" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e109 ">Model-based</td>

          </tr>
          <tr class="">
            <td data-label="Data Type" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e103 ">Passport</td>

            <td data-label="Pattern" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e106 ">
              <p>A number listed on a passport</p>
              <ul>
                <li>English: US and UK passport number formats are supported.</li>

                <li>French: France</li>

                <li>German: Germany</li>

                <li>Italian: Italy</li>

                <li>Japanese: Japan</li>

                <li>Spanish: Spain</li>

              </ul>
            </td>

            <td data-label="Detection Method" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e109 ">Regex pattern + relevant context words</td>

          </tr>
          <tr class="">
            <td data-label="Data Type" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e103 ">Phone Number</td>

            <td data-label="Pattern" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e106 ">
              <p>A phone number with or without a country code. Phone format is supported for a
                specific region and language.&nbsp;</p>
              <ul>
                <li>English-US: Area code (1 or 001) followed by a 10 digit number </li>

                <li>English-UK: Area code (44 or 044) followed by a 10 or 11 digit number </li>

                <li>French- France: Area code (33 or 033) followed by a 9 digit number </li>

                <li>German-Germany: Area code (49 or 049) followed by a 6 or 8 digit number </li>

                <li>Italian-Italy: Area code (39 or 034) followed by a 6 or 9 digit number</li>

                <li>Japanese-Japan: Area code (080) followed by a 10 or 11 digit number </li>

                <li>Spanish-Spain: Area code (34 or 034) followed by a 6 or 9 digit number </li>

              </ul>
            </td>

            <td data-label="Detection Method" class="slds-cell-wrap" style="vertical-align:top;" headers="d70589e109 ">Regex pattern + relevant context words</td>

          </tr>
        </tbody>
      </table>

These data types are supported in English (United States) only:

  
| Sensitive Data Type | Pattern | Detection Method |
| --- | --- | --- |
| United States driver’s license | US driving license number. Each state has a different format. The numbers range from 7 to 14 alphanumeric characters. | Regex pattern + relevant context words |
| United States Individual Taxpayer Identification Number | A US Individual Taxpayer Identification Number (ITIN) that includes 9 digits that start with a nine and contains 7 or 8 as the fourth digit. For example,  9XX-8X-XXXX | Regex pattern + relevant context words |
| United States Social Security number | A US Social Security Number (SSN) that includes 9 digits without any spaces, or 9 digits separated by dashes or spaces or periods. For example, 254941476, 254-94-1476, or 254.94.1476 | Regex pattern + relevant context words |

## Toxicity Detection

The Einstein Trust Layer detects toxicity in responses and assigns a score using machine learning models. This score is stored in Data 360 as part of the audit trail. To learn more, see [Toxicity Detection](https://help.salesforce.com/s/articleView?id=ai.generative_ai_trust_toxicity_detection.htm&release=264.0.0&type=5 "Einstein Trust Layer uses machine learning (ML) models to identify and flag toxic content in prompts and responses.").

   
| Language | Region / Locale | Language | Supported? |
| --- | --- | --- | --- |
| Danish | Denmark | `da_DK` | yes\* |
| Dutch | Belgium | `nl_BE` | yes\* |
| Dutch | Netherlands | `nl_NL` | yes\* |
| English | Australia | `en_AU` | yes\* |
| English | Canada | `en_CA` | yes\* |
| English | India | `en_IN` | yes\* |
| English | Ireland | `en_IE` | yes\* |
| English | Singapore | `en_SG` | yes\* |
| English | United Kingdom | `en_GB` | yes\* |
| English | United States | `en_US` | yes |
| French | Belgium | `fr_BE` | yes\* |
| French | Canada | `fr_CA` | yes\* |
| French | France | `fr_FR` | yes |
| French | Switzerland | `fr_CH` | yes\* |
| German | Austria | `de_AT` | yes\* |
| German | Germany | `de_DE` | yes |
| German | Switzerland | `de_CH` | yes\* |
| Italian | Italy | `it_IT` | yes |
| Italian | Switzerland | `it_CH` | yes\* |
| Japanese | Japan | `ja_JP` | yes |
| Korean | Korea | `ko_KR` | yes\* |
| Norwegian | Norway | `nb_NO` `nn_NO` | yes\* |
| Portuguese | Brazil | `pt_BR` | yes\* |
| Portuguese | Portugal | `pt_PT` | yes\* |
| Spanish | Mexico | `es_MX` | yes\* |
| Spanish | Spain | `es_ES` | yes |
| Swedish | Sweden | `sv_SE` | yes\* |
| Thai | Thailand | `th_TH` | yes\* |

yes\* = The presence of region-specific language presents difficulties in effectively identifying harmful or inappropriate content. Einstein Trust Layer doesn’t always identify such content with 100% accuracy, especially when it involves region-specific language patterns.

![Important](https://sf-zdocs-cdn-prod.zoominsoftware.com/tdta-ai-generative_ai-264-0-0-production-enus/f8ef6498-d7b2-44e3-9b34-ffd5a850aae9/images/icon_important.png)

Important Although our detection models have shown to be effective during internal testing, it's important to note that no model can guarantee 100% accuracy. In addition, cross-region and multinational use cases can affect the ability to detect specific data patterns. With trust as our priority, we're dedicated to the ongoing evaluation and refinement of our models.

## Prompt Injection Detection

Prompt injection detection is supported only in the English (United States) language.

## Citations

Citations are supported in English, French, German, Italian, Portuguese (Brazilian) and Spanish.