Device Type: desktop
Skip to Main Content Skip to Main Content

Salesforce Einstein and Customer Communication

This article was published on July 27, 2021

How can you use the data in your CRM to make your customer communication more effective with AI? If you're one of Salesforce's 150,000 customers, the answer lies in combining the Salesforce Einstein AI platform with the cloud communication APIs from Nexmo, the Vonage API Platform.

Data is the Lifeblood of 21st Century Business

Whether yours is a B2B or a B2C business, I'll bet that your CRM touches almost every one of your customer-facing processes. Today, platforms such as Salesforce combine sales and marketing enablement with a 360 degree customer view, easy integration with third-party systems, and data analytics.

That makes modern CRMs the veins through which a company's data flows. So, if data is the blood and the CRM is the veins, could AI be the brain?

Before we stretch this analogy too far, the answer is that AI needs data in just the same way as your real brain needs oxygen and energy. Salesforce's Einstein AI platform lets you combine the wealth of data in your Salesforce CRM with modern AI technologies, such as chatbots, prediction models, language parsing, and image recognition.

Using both the Einstein and Nexmo APIs, you can build this functionality directly into your customer communication strategy. Let's look at how.

What Does Einstein Offer?

Einstein's APIs let any developer build AI-powered services on top of data they provide or data stored in Salesforce. So, that means you can take advantage of Einstein's functionality even if you're not using the Salesforce CRM.

The Einstein Platform APIs fall into two categories:

  • natural language processing

  • image detection and classification.

So, how can they impact your customer communication strategy? As with most AI in 2019, it's all about making humans more effective.

Sentiment and Intent Analysis

Let's look at an example. You're a retailer who sells soft toys. One of your soft toys is an educational teddy bear that teaches children Spanish. The rest of your product line is cute and cuddly, but totally dumb.

As there's such a big difference between the two product lines you have two different customer care teams. One is trained in the needs of educational institutions and the other is trained to engage with the fantasy world that you've created around your cute and cuddly products.

Just to make things interesting, you've had some quality control issues recently and, to protect both your customers and your brand, you want to push unhappy customers directly to the front of the queue.

How can Nexmo with Salesforce Einstein help you here?

As part of an omnichannel customer communication strategy, your customers should be able to reach you through whichever method suits them best.

Let's say that a customer contacts you through WhatsApp. Now, should you have separate WhatsApp accounts for sales and support? No, because while that might make things easier for you it adds unnecessary complexity for your customers.

Instead, using Nexmo's Messages API you connect your WhatsApp Business account to your Salesforce CRM and use your customer data, with Einstein APIs, to decide automatically on the best course of action.

The flow might look something like this:

Without human intervention, inbound messages that are clearly about one topic or the other are assigned to the correct team.

It's a similar story for judging whether the customer is unhappy or not.

Once a message enters your system, via Nexmo's APIs, you can then pass it to the Einstein natural language processing APIs for automatic categorization and prioritization.

Image Classification

In our cuddly toy example, let's imagine that two different factories produce the language teaching toy. Factory A has no quality control problems, whereas 20 percent of Factory B's output has been faulty.

You have a "no questions asked" refund policy for Factory B toys but you don't want to tell the world. Instead, you want to identify queries that pertain to Factory B toys quickly and then offer a refund to those customers only.

To make your life easier, the base of Factory B toys has a green triangle stamped on the battery cover and Factory A toys have a blue square.

Rather than immediately passing unhappy customers to a human agent, you could instead use a chatbot to ask the customer to take a photo of the battery cover and share it using WhatsApp.

Once you have the image, you can then pass it to Einstein's image classification API. Now, this requires some prep: you'll need to have trained Einstein to recognize the green triangles versus the blue squares. However, following that training the API will tell you which photos have a green triangle and which have a blue square.

You can then use the chatbot to automatically offer owners of green triangle (or Factory B) toys a refund. That way, the customers get quick satisfaction and your human agents can work on less straightforward queries.

Putting It Into Action

If you're a Salesforce customer, your development team can start working with the Einstein APIs right away. The combination of Nexmo's APIs and the Einstein AI platform gives you a low-cost way to deploy artificial intelligence in your customer communication strategy.

And if you want to learn more about the role of artificial intelligence in your customer communication strategy then read our primer on AI in the contact center.

Written by Vonage Staff

Deskphone with Vonage logo