This short YouTube demo explains how to use AI-powered semantic search with Salesforce Data Cloud and Einstein to find and surface similar customer cases, improving agent efficiency and customer satisfaction. The speaker contrasts semantic search with brittle keyword search, demonstrates a live walkthrough (using Einstein Copilot to summarize a “scanner wand is broken” case and opening related cases), and outlines the implementation steps: choose an embedding model, convert case data to vectors, store/index them in a vector DB, expose a search API, and keep vectors synchronized. The video concludes by emphasizing business benefits (reduced attrition, faster onboarding, higher productivity and customer satisfaction) and links to more resources.
Chapters
Today I'm going to show you how to use AI to find similar cases with Data Cloud's vector database and semantic search features. Now, you're probably thinking, great, but what does that have to do with creating better customer service? Well, it's simple. When an agent can quickly retrieve and review similar cases to see how they were resolved, they'll be more confident and efficient as they continue to learn the company, and they'll be able to resolve complex issues where they There might not be good articles that already exist in your knowledge base, but how are your agents going to find cases that might be related to their current case? Now you might be thinking keyword search, right? After all, that's what we're used to. But keyword searches are limited. For example, if a customer comes to you with I love Northern Trail Outfitter boots, I bought many pairs, it's very clear what the customer is looking for, but a simple command F search function is going to leave you high and dry. Now, what would you even search for to get a quick and relevant answer? It would be way easier to say, show me anything related to this message. Which means you need a search that understands the meaning or semantics of the message. You need a semantic search because it's much more impactful to use a system that understands the common language and can pull relevant cases. That'll significantly improve the amount of time it takes to resolve your case while providing the most relevant information possible to your customer. Now here, let me show you how this all works. We're going to start by logging in so we can receive email to Case. Now let's take a look at the scanner wand is broken case. Now one of the things you'll notice here is if we look at the knowledge area, we aren't seeing any Einstein article recommendations. That's a visual cue that maybe this is something that doesn't have a knowledge article. Side note, if you want to see how Einstein can help make your knowledge base work harder for you, I dropped a video down in the description and trust me, it's a good one. Alright, back to our broken scanner wand. Now let's start by asking Einstein Copilot, our conversational AI assistant for CRM, to give us a summary of the case. This step is important because if we're going to do a semantic search to find similar cases, we need Einstein to start with an overall understanding of the case. Okay, so Einstein found some similar cases. One that's still open and two that are closed. Let's take a look at these cases by clicking on the case number to open a tab for each. Now, for this specific case, I can see in the case summary that disconnecting and then reconnecting the scanner wasn't helpful, but changing the batteries and placing the scanner back in the cradle did the trick. I can also quickly look at the other two cases to see to see how they were resolved too. And boom. In just a few easy steps, with the help of Einstein, I have the info I need to resolve this issue for my customer. Now, if you're wondering how to get started using your case data for semantic searches, look, all you have to do is select an embedding model, prepare the data, convert that data into vectors, store those vectors in a vector database, index everything, and then create an API for handling the semantic searches. Oh, and also keep refreshing the vector database as your data changes. Sounds complicated, right? But it's not. Because here inside Data Cloud, with a few clicks, your team can connect data from anywhere in any format, which includes case data that you'll want to enable for semantic search. Now, after it's connected, it just takes a few more clicks to set up a search index for this data, which automatically handles everything that the embedding model, the conversion to vectors, the storage of those vectors, the API, and the continual data synchronization. Which means Einstein is ready to help your service agents. Now, semantic search capabilities are obviously great for service teams that want to build up their knowledge base, but its real value is in making your agents jobs easier. And that helps you reduce attrition, speed up onboarding, increase productivity, and best of all, increase customer satisfaction. All right, to learn more about how Salesforce AI can help you and your business increase productivity, margins and customer satisfaction, click the link on your screen right now. Also, I'll drop a few resources in the description because that's what I do. Thanks so much for watching and we'll see you next time.
