Introduction
In this blog, I am going to explain how to use salesforce Einstein Intent. Einstein Intent Categorize unstructured text into user-defined labels to better understand what users are trying to accomplish. You can use Einstein Intent API to analyze text from emails, chats, or web forms and many other use cases like
- Determine what products prospects are interested in and send customer inquiries to the appropriate salesperson.
- Identify topics across unstructured text in emails, meeting notes, or account notes to summarize key points.
- Route service cases to the correct agents or departments, or provide self-service options.
- Understand customer posts to provide personalized self-service in your communities
Prerequisites
Salesforce Einstein Platform Services APIs user OAuth token for authenticating. To generate a token, you create a JWT payload, sign the payload with your private key, and then call the API to get the token. So you need to get your own einstein_platform private key. please follow this steps to get your einstein_platform key.
- Go to the sign up page.
- Click Sign Up Using Salesforce.
- On the Salesforce login page, type your username and password, and click Log In.
- Click Allow so the page can access basic information, such as your email address, and perform requests.
- Download Key to save the key locally. The key file is
einstein_platform.pem - Upload the einstein_platform.pem key to Salesforce files.
- Copy JWTBearerFlow and JWT Apex Class from this git repository https://github.com/salesforceidentity/jwt
Refer this link for more information on the above steps.
Create custom metadata type or custom setting whatever works for you here to store the service endpoint, token endpoint and account details. Below is the custom metadata type which we are going to use it in this blog post.
Let’s Get Started.
As we are going to user cURL for model upload and training, we need to Generate a token for authentication. Go to this link and generate the token as shown below. We are going to use this generated token for model upload and train the model for predictions.
Remote Site Setting
Add remote site setting as shown below.
Step 1: DEFINE YOUR CLASSES AND GATHER DATA
In this step, you define the labels that you want the model to output when text is sent into the model for prediction. Then you gather text data for each of those labels, and that text is used to create a model.This is typically the most time-consuming part of the process. To make it easier for you to go through these steps, we provide a case routing .csv file that you can use.
The labels in the case routing dataset define the intent behind the text. The intent can then be used to route that case to the right department. Those labels are:
- Billing
- Order Change
- Password Help
- Sales Opportunity
- Shipping Info
STEP 2: CREATE THE DATASET
In this step, you use the data you gathered to create a dataset. In the following command, replace with<TOKEN> your JWT token and run the command. This command:
- Creates a dataset called
case_routing_intentfrom the specified .csv file by accessing the file via a URL - Creates five labels as specified in the .csv file
curl -X POST -H "Authorization: Bearer 7KQH5DBEZNE2Z3H5L55I4POCYTAB3GMTSDA6HV264XGDGDSQGJACCO7M4DF5DTDGTDJI37WR4NGEPNFJZCGRGS2IBDRWF5OHIXF3SPY " -H "Cache-Control: no-cache" -H "Content-Type: multipart/form-data" -F "data=@C:\Users\rvakati\Desktop\New folder\SFDC Intent\case_routing_intent.csv" -F "type=text-intent" https://api.einstein.ai/v2/language/datasets/upload
{
"id": 1035176,
"name": "case_routing_intent.csv",
"createdAt": "2017-06-22T19:31:58.000+0000.",
"updatedAt": "2017-06-22T19:31:58.000+0000",
"labelSummary": {
"labels": []
},
"totalExamples": 0,
"available": false,
"statusMsg": "UPLOADING",
"type": "text-intent",
"object": "dataset"
}
To verify that the data has been loaded, make a call to get the dataset. Replace with<TOKEN> your JWT token and with<DATASET_ID> ID of the dataset you created.
curl -X GET -H "Authorization: Bearer 7KQH5DBEZNE2Z3H5L55I4POCYTAB3GMTSDA6HV264XGDGDSQGJACCO7M4DF5DTDGTDJI37WR4NGEPNFJZCGRGS2IBDRWF5OHIXF3SPY" -H "Cache-Control: no-cache" https://api.einstein.ai/v2/language/datasets/1035176
Once you run the above command you will get below response with each label in your dataset.
{"id":1035176,"name":"case_routing_intent","createdAt":"2018-01-22T06:07:25.000+0000","updatedAt":"2018-01-22T06:07:26.000+0000","labelSummary":{"labels":[{"id":319142,"datasetId":1035176,"name":"Order Change","numExamples":26},{"id":319143,"datasetId":1035176,"name":"Sales Opportunity","numExamples":44},{"id":319144,"datasetId":1035176,"name":"Billing","numExamples":24},{"id":319145,"datasetId":1035176,"name":"Shipping Info","numExamples":30},{"id":319146,"datasetId":1035176,"name":"Password Help","numExamples":26}]},"totalExamples":150,"totalLabels":5,"available":true,"statusMsg":"SUCCEEDED","type":"text-intent","object":"dataset"}
STEP 3: TRAIN THE DATASET
Now you need to train your data set to create prediction ready model. Use this command to train the dataset and create a model. Replace <TOKEN> with your JWT token and <DATASET_ID> with ID of the dataset you created
curl -X POST -H "Authorization: Bearer 7KQH5DBEZNE2Z3H5L55I4POCYTAB3GMTSDA6HV264XGDGDSQGJACCO7M4DF5DTDGTDJI37WR4NGEPNFJZCGRGS2IBDRWF5OHIXF3SPY" -H "Cache-Control: no-cache" -H "Content-Type: multipart/form-data" -F "name=Case Routing Model" -F "datasetId=1035176" https://api.einstein.ai/v2/language/train
You will get below response from the above service.You will get modelId which we will use it for further predictions.
{"datasetId":1035176,"datasetVersionId":0,"name":"Case Routing Model","status":"QUEUED","progress":0,"createdAt":"2018-01-22T06:16:14.000+0000","updatedAt":"2018-01-22T06:16:14.000+0000","learningRate":0.0,"epochs":0,"queuePosition":1,"object":"training","modelId":"PF4WICRUGR2BOO7QJ5MESD56MU","trainParams":null,"trainStats":null,"modelType":"text-intent"}
Use the modelId to make this call and get the training status. Replace<TOKEN> with your JWT token and withMODEL_ID the ID of the model you created.
curl -X GET -H "Authorization: Bearer 7KQH5DBEZNE2Z3H5L55I4POCYTAB3GMTSDA6HV264XGDGDSQGJACCO7M4DF5DTDGTDJI37WR4NGEPNFJZCGRGS2IBDRWF5OHIXF3SPY" -H "Cache-Control: no-cache" https://api.einstein.ai/v2/language/train/PF4WICRUGR2BOO7QJ5MESD56MU
{"datasetId":1035176,"datasetVersionId":17909,"name":"Case Routing Model","status":"RUNNING","progress":0.01,"createdAt":"2018-01-22T06:16:14.000+0000","updatedAt":"2018-01-22T06:17:17.000+0000","learningRate":0.0,"epochs":1000,"object":"training","modelId":"PF4WICRUGR2BOO7QJ5MESD56MU","trainParams":null,"trainStats":null,"modelType":"text-intent"}
STEP 4: SEND TEXT IN FOR PREDICTION
Now your model is ready to go! To test it out, send some text in for prediction. This cURL call takes the ofmodelId the model from which you want to return a prediction and the text string to analyze. Replace with<TOKEN> your JWT token and with<MODEL_ID> the ID of your model.
curl -X POST -H "Authorization: Bearer 7KQH5DBEZNE2Z3H5L55I4POCYTAB3GMTSDA6HV264XGDGDSQGJACCO7M4DF5DTDGTDJI37WR4NGEPNFJZCGRGS2IBDRWF5OHIXF3SPY" -H "Cache-Control: no-cache" -H "Content-Type: multipart/form-data" -F "modelId=PF4WICRUGR2BOO7QJ5MESD56MU" -F "document=how is my package being shipped?" https://api.einstein.ai/v2/language/intent
The response looks like this JSON. The model predicts that the text indicates that the user has a comment or question about shipping, so the model returns asShipping Info the top probability. Your app can then use this information to route the case to the right department or agent.
{
"probabilities": [
{
"label": "Shipping Info",
"probability": 0.82365495
},
{
"label": "Sales Opportunity",
"probability": 0.12523715
},
{
"label": "Billing",
"probability": 0.0487557
},
{
"label": "Order Change",
"probability": 0.0021365683
},
{
"label": "Password Help",
"probability": 0.0002156619
}
],
"object": "predictresponse"
}
Code :
public class EinsteinIntent {
@AuraEnabled
public static List<Probabilities> getIntent(String sentimentModel , String textforIntent){
ContentVersion base64Content = [SELECT Title, VersionData FROM ContentVersion WHERE Title = 'einstein_platform' LIMIT 1 ];
Einstein_Settings__mdt einsteinSettings = [Select DeveloperName, Label , Account_Email__c , Service_EndPoint__c , Token_EndPoint__c
from Einstein_Settings__mdt where DeveloperName ='Intent' Limit 1] ;
DateTime tokenExpireTime = DateTime.now().addMinutes(30);
String tokenExpireTimeinUnixFormate = ''+tokenExpireTime.getTime()/1000;
String keyContents = base64Content.VersionData.tostring();
keyContents = keyContents.replace('-----BEGIN RSA PRIVATE KEY-----', '');
keyContents = keyContents.replace('-----END RSA PRIVATE KEY-----', '');
keyContents = keyContents.replace('\n', '');
JWT jwt = new JWT('RS256');
jwt.pkcs8 = keyContents;
jwt.iss = 'developer.force.com';
jwt.sub = einsteinSettings.Account_Email__c ;
jwt.aud = einsteinSettings.Token_EndPoint__c;
jwt.exp = tokenExpireTimeinUnixFormate;
String access_token = JWTBearerFlow.getAccessToken(einsteinSettings.Token_EndPoint__c, jwt);
Http http = new Http();
HttpRequest req = new HttpRequest();
req.setMethod('POST');
req.setEndpoint(einsteinSettings.Service_EndPoint__c);
req.setHeader('Authorization', 'Bearer ' + access_token);
req.setHeader('Content-type', 'application/json');
String body = '{\"modelId\":\"'+ sentimentModel + '\",\"document\":\"' + textforIntent + '\"}';
req.setBody(body);
HTTPResponse res = http.send(req);
System.debug('res'+res.getBody());
JSONParser parser = JSON.createParser(res.getBody()) ;
String label ='';
Decimal probability = 0 ;
List<Probabilities> probabilities = new List<Probabilities>() ;
while (parser.nextToken() != JSONToken.END_OBJECT) {
if (parser.getCurrentToken() == JSONToken.FIELD_NAME) {
String text = parser.getText();
if (parser.nextToken() != JSONToken.VALUE_NULL) {
if (text == 'probabilities') {
probabilities = new List<Probabilities>();
while (parser.nextToken() != JSONToken.END_ARRAY) {
probabilities.add(new Probabilities(parser));
}
}
}
}
}
return probabilities ;
}
public class Probabilities {
@AuraEnabled
public String label { get; set; }
@AuraEnabled
public Double probability { get; set; }
public Probabilities(JSONParser parser) {
while (parser.nextToken() != JSONToken.END_OBJECT) {
if (parser.getCurrentToken() == JSONToken.FIELD_NAME) {
String text = parser.getText();
if (parser.nextToken() != JSONToken.VALUE_NULL) {
if (text == 'label') {
label = parser.getText();
} else if (text == 'probability') {
probability = parser.getDoubleValue();
}
}
}
}
}
}
}
Lightning component
<aura:component implements="flexipage:availableForAllPageTypes,flexipage:availableForRecordHome,force:hasRecordId"
access="global" controller="EinsteinIntent">
<aura:attribute name="modelName" type="String" default="PF4WICRUGR2BOO7QJ5MESD56MU"/>
<aura:attribute name="textToAnlyze" type="String" default="Can I change my credit card"/>
<aura:attribute name="predectionResult" type="List[]"/>
<form class="slds-form--stacked">
<lightning:input aura:id="formInp" label="Enter text"
name="Intent Text"
value="{!v.textToAnlyze}"
required="true"/>
<lightning:button label="Predict "
class="slds-m-top--medium"
variant="brand"
onclick="{!c.Predict}"/>
</form>
<table class="slds-table slds-table--bordered slds-table--striped" style="margin-top:30px;">
<thead>
<tr>
<th>Label</th>
<th>Probability</th>
</tr>
</thead>
<tbody>
<aura:iteration items="{!v.predectionResult}" var="predectionResultVal" >
<tr>
<td>{!predectionResultVal.label}</td>
<td>{!predectionResultVal.probability}</td>
</tr>
</aura:iteration>
</tbody>
</table>
</aura:component>
({
Predict : function(component, event, helper) {
helper.load(component);
}
})
({
load: function(component) {
var action = component.get("c.getIntent");
action.setParams({
"sentimentModel": component.get("v.modelName") ,
"textforIntent": component.get("v.textToAnlyze")
});
action.setCallback(this, function(response) {
var state = response.getState();
console.log('state'+state);
if (state === "SUCCESS"){
var predectionResult = response.getReturnValue();
console.log(predectionResult);
component.set("v.predectionResult" , predectionResult) ;
} else if(state == "ERROR"){
var errors = response.getError();
console.log(errors);
}
});
$A.enqueueAction(action);
},
})
You can refer this link for additional information .
Beyond Cloud 