Start Using Article Recommendations
Learning Objectives
After completing this unit, you’ll be able to:
- Describe the difference between Suggested Articles and Einstein Article Recommendations.
- Select the right fields for Einstein to learn from.
- Use relevance and confidence scoring to help Einstein improve.
What Makes Einstein Article Recommendations So Powerful?
As Salesforce veterans, Sita and her team already use a tool called Suggested Articles. Suggested Articles is a Service Cloud feature that recommends knowledge articles to your agents. However, Suggested Articles is a keyword-based search that can’t learn from your case data. Einstein Article Recommendations, on the other hand, uses data from previous cases to produce more accurate recommendations in a matter of seconds. And unlike Suggested Articles, it can refine suggestions.
If you start using Einstein Article Recommendations, we recommend disabling Suggested Articles to ensure the best user experience for your support team. Otherwise, agents see two sets of articles that relate to the case.
Get Started with Einstein Article Recommendations
Sita sees that Einstein Article Recommendations has significant case-solving potential, so she decides to try it out. Here are the steps of her adoption journey.
Like Sita, to adopt Einstein Article Recommendations, review the requirements and make sure it’s available in your org. In Einstein fashion, it’s automatically available with the Lightning Knowledge Console in Enterprise, Performance, or Unlimited editions. If you have Lightning Knowledge and one of these editions, you already have Einstein Article Recommendations in your org—you just need to roll it out! We cover that next.
Help Einstein Navigate Your Data
Einstein Article Recommendations is a tool of the future that’s ready today. That means the days of agents manually searching for and attaching articles can be over. To do this, Einstein gathers data from your closed cases, especially cases with articles attached, and from your knowledge base. To get up and running, you select case fields and article fields for Einstein to learn from. Think of this as giving Einstein a map to navigate through your data. The better the map, the more accurate Einstein’s recommendations.
In this module, we assume you're an Einstein Article Recommendations admin with the proper permissions to set up the feature. If you’re not an admin for Einstein, that’s OK. Read along to learn how your admin takes the steps in a production org. Don’t try to follow these steps in your Trailhead Playground. Einstein Article Recommendations isn’t available in the Trailhead Playground.
Steps to selecting fields for Einstein:
- From Setup, in the Quick Find box, enter
Einstein Article Recommendation, and select Einstein Article Recommendations.
- Click the toggle to enable the feature.
- To create your model, under Select Fields, click Select | Next.
- Select the supported languages that you want to include in your model, and click Next.
Einstein uses a single model to generate article recommendations in Dutch, English, French, German, Italian, Portuguese, and Spanish. Verify that your selected languages are also active in your Knowledge settings.
- Select the fields from cases that you want to incorporate into your Einstein model, and click Next.
5. Select the fields from Knowledge articles that you want to incorporate into your Einstein model, and click Save.
How do you know you’re selecting the best case and knowledge article fields? Here are some best practices.
-
Choose data-rich case and article fields. Look for text fields that are rarely left blank and typically contain multiple words that tell you what the case or article is about. Unpopulated or uninformative fields make it hard for Einstein to make accurate recommendations.
-
Rank fields based on importance. Einstein wants to know what matters most to you. By ranking your fields in a case, you’re essentially telling Einstein what to look for first.
- Choose fields that might be updated. When a case is updated, new article recommendations appear. To show agents the most relevant articles, select fields that are likely to change over a case’s lifecycle.
If you want to change which fields Einstein learns from, you can always return to this Setup page, update your fields, and rebuild your model.
After you select fields, there’s a couple more steps for you to take: Build and activate your Einstein model and give agents access to Einstein recommendations. We don’t cover these steps in detail in this badge, but check out the Resources section at the end of this unit for more information.
Gain Confidence with Relevance Scoring
Einstein also gives you and your support team data to measure success. The relevance score for Einstein Article Recommendations tells your agents how relevant an article is to a customer case. This relevancy is expressed as a percentage and appears above the article’s title.
An agent interacts with an article recommendation by hovering over, accepting, or dismissing a recommendation by clicking Not Helpful. Einstein records these actions to challenge an article’s relevance score over time and provide more accurate recommendations in the future.