Einstein article recommendations trains Salesforce<\/strong> via its intelligent engine to recommend the right articles for the right cases. The engine can do this thanks to a set of case and knowledge fields that you choose to let the matching happen. In other words, the recommendation engine will match the key words for certain cases with key words for certain knowledge articles. Then it will be smart enough to suggest the same articles in the future.<\/p>\n But then you might ask yourself :<\/p>\n How is it different from the classic Salesforce article recommendation ?<\/strong><\/p><\/blockquote>\n That\u2019s a very good question. And the answer lies within one word \u201ctraining<\/strong>\u201d.<\/p>\n You see, the classic method of recommendation is based solely on the key words and the similarities between the case and knowledge fields but with Einstein, the system trains itself. This means that it is not required to have an exact match between the different words. The engine, if well trained, will do the matching on its own and without the need to have the exact wording.<\/p>\n Let\u2019s take the following example : <\/strong><\/p>\n You\u2019re a car repair company and your agents are receiving emails about engine problems. The experienced agents with all the knowledge they have know that the first thing that they must do is to ask the customer to change the oil by sending them a knowledge article on how to do so.<\/p>\n In the meantime, the Einstein engine is learning to associate the sentence \u201cengine problem\u201d from the case object with the sentence \u201cchange oil\u201d from the article title. Then in the future, Salesforce will automatically suggest this article to any case of the type \u201cengine problem\u201d so that the new inexperienced agents will be able to help customers efficiently.<\/p>\n Ok, so that all is fine and good but :<\/p>\n How can we get into business and activate this thing ?<\/strong><\/p><\/blockquote>\n From Setup, in the Quick Find box, enter Einstein Article Recommendations, and select Einstein Article Recommendations<\/strong>.<\/p>\n Under Select Fields, click Select | Next.<\/p>\n Select the fields from cases that you want to incorporate into your model, and click Next.<\/p>\n <\/p>\n Then we do the same thing but with the articles fields.<\/p>\n <\/p>\n From Setup, in the Quick Find box, enter Einstein Article Recommendations, and select Einstein Article Recommendations. Confirm that case and knowledge article fields are selected for your model. Then, under Build Model, click Build | Build Model.<\/p>\n The model building can take about 48 hours so no worries here !<\/p>\n After your model finishes building, click Activate to start showing recommendations to agents.<\/p>\n And then, our final step is to show the recommendations to our users. To do that you can choose whether you want to give access to all Lightning users or only a specific profiles or permission sets.<\/p>\n Don\u2019t forget to give those profiles or permission sets the \u201cView and Act on Einstein Article Recommendations\u201d right.<\/p>\n So we\u2019re all set and done !<\/p>\n If we did everything right, the users will start to see the recommended articles in the service cloud console with a relevant score as well for the prediction accuracy.<\/p>\n And voil\u00e0 ! The Salesforce Einstein article recommendation is up and running. Congratulations !<\/p>\n Remember the fifth of November! Besides, remember that to have a successful experience you\u2019ve got to have at least three or more articles in your base which is a requirement. Salesforce also recommends having at least 1000 or more cases. And good news starting soon new seven languages will be powered by this feature and not only English as was the case this far.<\/p>\n Another great advice is to take care of the quality of your articles in terms of titles, types and content. Not only you\u2019ll help Einstein making better decisions but your agents will thank you anyway. Garbage in garbage out!<\/p>\nHow can I configure it ?<\/strong><\/h2>\n
To create your model :<\/h3>\n
![\"Set](\"https:\/\/texei.com\/dev\/wp-content\/uploads\/2021\/01\/0_ETX5D5Wnin1_pYBm.png\")
<\/figure>\n![\"Set](\"https:\/\/texei.com\/dev\/wp-content\/uploads\/2021\/01\/0_bhwFda-CX2V9fY80.png\")
<\/figure>\nThen to activate the model :<\/h3>\n
![\"click](\"https:\/\/texei.com\/dev\/wp-content\/uploads\/2021\/01\/0_hFRaG4cGNDoSTKEt.png\")
<\/figure>\nHow can the users see the recommendations ?<\/strong><\/h2>\n
![\"\"](\"https:\/\/texei.com\/dev\/wp-content\/uploads\/2021\/01\/0_PgHMFLvHxiij4qQr.png\")
<\/figure>\nTips<\/strong><\/h2>\n