Improve Einstein's Recommendations

Learning Objectives 

After completing this unit, you’ll be able to:

  • Update Einstein Article Recommendations as needed.
  • Describe best practices for Einstein Article Recommendations.

Put Einstein to the Test

Maria, Ursa Major Solar’s admin, is more than ready to see Einstein Article Recommendations in action. Her specially selected agents are now able to see article recommendations on cases in the console. Let’s follow Maria and one of her agents on their Einstein Article Recommendations journey.

Give Einstein Feedback

Maria notices her first case coming in. Her agent opens the case. Several article recommendations appear in the Knowledge component. The articles are listed in order of their relevance to the case.

Here’s a case example of what her recommendations can look like.

Einstein Article Recommendations example with dropdown options shown: Edit as Draft, Archive, and Attach Article.

Maria prompts her agent to attach the first article, which she knows to be helpful for understanding this new product. The agent clicks Attach Article. The agent notices a not-so-helpful article recommendation right below, so Maria asks the agent to click Not Helpful.

As mentioned in the previous unit, Einstein records your agent’s actions each time they interact with the recommended articles. The more your agents interact with a recommended article to dismiss or accept a recommendation, the more Einstein learns. Together, your agents and Einstein can provide a better customer experience.

Improving Your Model

As your company grows, your customers’ inquiries can evolve. To keep up Einstein’s article recommendations, here are a few tips to keep your agents primed.

  • Keep attaching articles to cases. We recommend that agents attach an article to a case if they referred it to while working on the case.

  • Keep expanding your knowledge base. The more articles in circulation, the more Einstein has to attach and learn from. Einstein can suggest new articles within a day of publication.

  • Encourage agents to dismiss recommendations that aren’t helpful. Einstein considers dismissals when formulating recommendations.

  • Update your fields. Have your support teams continue filling out case fields consistently, and periodically reconsider which fields are included in the model.

  • Review the model scorecard. A scorecard summarizes model effectiveness and data quality. The scorecard’s metrics can help you identify where there are opportunities to refine your data. 

The Einstein Article Recommendations tool is designed to make your agents’ lives easier and help your company develop better relationships with your customers. Undoubtedly, this next-generation technology has helped Ursa Major Solar close more cases with its amazing products and equally amazing support team.

Einstein Article Recommendations is also just one of the many tools offered by Einstein for Service. Close more cases faster, easier, and smarter by becoming familiar with the rest of the Einstein for Service features, and learning more about how these features can work together to improve your business.


Continue a aprender de graça!
Inscreva-se em uma conta para continuar.
O que você ganha com isso?
  • Receba recomendações personalizadas para suas metas de carreira
  • Pratique suas habilidades com desafios práticos e testes
  • Monitore e compartilhe seu progresso com os empregadores
  • Conecte-se a orientação e oportunidades de carreira