en flag +1 214 306 68 37

Salesforce Artificial Intelligence for Marketing, Sales and Service

CRM Expert and IT Architect


Editor's note: We look into a Salesforce AI tool called Einstein and teach you to use all its capabilities to the fullest. And if you seriously consider a possibility to implement an AI-powered solution, we offer to you professional AI consulting services and full guidance along this journey.

Not Terminator, not Chappie, not Ava and not even JARVIS. What movies are telling us about artificial intelligence and what AI really means for marketing, sales and service are two completely different things. First, AI doesn’t have a body. Second, it doesn’t engage in fights. Third, it doesn’t even speak. So, what actually is AI for business? Technically, a set of machine learning and deep learning capabilities that analyzes data and provides easy-to-follow predictions and recommendations to make your employees’ decisions more informed and rapid.

As to Salesforce AI, such a set of capabilities is embodied in an ultimate ‘helper’ called Einstein. It gives you the needed insights at the needed moment and strives to help your employees work at their most efficient within all Salesforce, be it Marketing Cloud, Service Cloud, Sales Cloud or anything else. For example, it has a lead scoring possibility. Based on historical data analysis and customer segmentation, Einstein evaluates how likely a certain lead is to convert allowing sales reps to see which leads to focus on and which ones to ‘let go.’

But why is Salesforce Artificial Intelligence so hot and what else can it do? Let’s investigate!

Salesforce AI

Why so much fuss?

Einstein takes all your historical data, analyzes it against pre-established sets of parameters, creates data models, which then are trained and tested on huge data sets. This is how you get your recommendations and predictions. Einstein just tells you what came next in a similar situation or what action influenced the matter well in such a case in the past. But, having once created a model, Einstein doesn’t just use it till the end of times. It runs the model on all the incoming data to see whether it is still adequate and whether it needs changing. This approach allows Einstein to make more accurate predictions and recommendations. Essentially, working with Salesforce is like using software with its own genius data scientist working 24/7 and never getting tired.

Besides, to ensure trust to Einstein, Salesforce AI provides explanations to all its statements. For instance, it shows top factors that influence each score Einstein gives.

But there’s a ‘but’

To use Einstein, you need to have a huge lot of data for him to learn from. And if we’re talking about lead scoring, Einstein will need over 1,000 leads created in the system plus about 120 opportunities converted to deals, all within the past 6 months. So, you need to be using Salesforce for quite some time and at quite a ‘level’ to be able to make use of Einstein. Evaluating your readiness to such a novelty can be quite difficult, and that is why the platform offers an Einstein Readiness Assessor, which is quite handy.

What can it do?

1. Einstein + Marketing

Salesforce AI can help your marketers:

  • Perform engagement scoring and prepare email content

Based on the characteristics of each particular customer, Einstein can see how likely they are to engage with your content (look through offers, follow links, purchase something). And based on their preferences, Einstein can recommend personalized email content and optimal send times. Just like with lead scoring, this allows your marketers to spend their time much more productively.

Email engagement prediction with Salesforce Einstein

Email engagement prediction with Salesforce Einstein

Making it your own. Even a powerful thing like Einstein needs customization. Here, you can have your own rules set for Einstein’s content recommendations. For instance, never to show particular groups of products to particular customer segments, because your company’s experience proves they do not go well together. Or you can have different engagement scoring factors set up for each customer segment.

  • Track posts on social media

Due to the ability of post classification, Einstein sorts social media posts that mention your company and/or products into particular categories. If posts require it (their sentiment is negative), Einstein tasks certain teams with replying to them and solving the customers’ problems, which helps to maintain positive company image. Besides that, Einstein is trained to filter out spam. And Einstein Vision, with its image recognition and classification, spots additional mentions related to your company through photos and pictures, even when no words signalize that. It helps to track the audiences’ interest in your products as well as your brand representation in the media.

Tracking social media with Salesforce Einstein

Tracking social media with Salesforce Einstein

Making it your own. Here, Einstein is particularly good. But to make it work even better, you may need to brush things up a bit. First, you can have your post categories set up, just like the corresponding procedure of directing posts to particular teams for response and resolution. Second, you can ask your Salesforce consulting team to alter sentiment-related rules to, say, tailor them to certain customers whose tone is always more aggressive and negative than that of others. And as to image classification, your consultants can ‘teach’ Einstein to filter your employees’ posts into a separate category to better work with your HR brand representation.

2. Einstein + Sales

Salesforce AI can help sales reps to:

  • Prioritize leads and opportunities better

Einstein’s predictive abilities don’t stop with engagement and lead scoring (which are already cool). Based on a similar model, Einstein can do opportunity scoring. It means that it can predict whether an opportunity is likely to close the deal with your company, which adds more value to your sales reps’ opportunity nurturing possibilities. To top it all, both in case with leads and opportunities, Einstein gives ‘relationship’ recommendations on beneficial next steps. For example, if one of your numerous opportunities has reengaged, Einstein will tell you that you need to call them or send an email. And, as a bonus, it synchronizes your email and calendar with Salesforce and eliminates data entry to leave your sales reps more time to deal with more arduous matters.

Lead scoring with Salesforce Einstein

Lead scoring with Salesforce Einstein

Making it your own. Before getting overexcited with scoring, you should analyze what particular factors influence your lead/opportunity conversion. And after laying down the factors, you can have the scoring process tailored to your business. Besides, you can have your platform set up to import the leads and opportunities that score too little into a separate database to further investigate how such cases’ conversion rate can be reinforced.

  • Track account news and look for changes in account details

Einstein understands that not only usual historical data can influence deals but also various structural and strategic changes within companies. To keep an eye out for them, it browses available external data sources to look for any cues of mergers, acquisitions, expansions or leadership changes within your prospects’ companies. Or analyzes recent emails and warns sales reps about their contacts’ mentioning competitors or changing jobs. As a bonus, Einstein looks through emails and creates new contacts in Salesforce.

Tracking account changes with Salesforce Einstein

Tracking account changes with Salesforce Einstein

Making it your own. You may have specific cases where someone in the leadership of a particular company is blocking your deal. In this case, you could benefit from, say, setting up an alert to go off when the leadership of that company changes.

3. Einstein + Service

Salesforce AI can help your service agents and managers to:

  • Manage customer satisfaction rates

In real time, Einstein analyzes agent availability, wait times and queues to predict customer satisfaction levels. If those are alarming, service managers are warned and provided with recommendations on how to mend the situation. Besides, Einstein recommends variants of response to customer inquiries and automatically classifies cases, which contributes to speedy case resolution and positively affects customer satisfaction rates.

Automated Case Classification and Recommended Responses in Service Cloud Einstein

Automated case classification and recommended responses with Salesforce Einstein

Making it your own. To smooth things with wait times, you may want to set up your Einstein to play/show the ‘queueing’ customers a few personalized offers. And if it is still not enough to stall them, then, offer a discount. Also, you may want to have Einstein’s service management recommendations tuned in accordance with your business.

  • Redirect calls and keep agents informed

Before a customer even talks to an agent, Einstein can analyze the information on the case, classify it and redirect the calling customer to the needed agent. On the other end of this process, when Einstein redirects a call, it provides the service agent with all needed info on the case: customer history, service issue, etc.

Making it your own. You may decide that your agents will need extra info to solve cases. Or that each case type will require a separate set of details shown/sent to the agent. Or you may want to ‘tell’ Einstein what parameters it needs to analyze while classifying cases and choosing agents to refer to. Anyway, tailoring Service Cloud Einstein to your specific view is more than possible.

What’s next?

Salesforce is constantly changing, with new features being added or improved as often as 3 times a year (in its Winter, Spring and Summer releases). Einstein is no exception. Currently, some of its capabilities are in beta, while some others are only scheduled for release.

In beta

As of April 2018, the biggest ones here are Einstein Language, Object Detection and Einstein Forecasting. All of them are working right now but are still being ‘polished.’

  1. Einstein Language has two parts: Sentiment and Intent API. The former analyzes the tone of a message (negative, neutral and positive), while the latter understands what the author of the message needs (for instance, to know delivery time). Einstein Language is used in marketing and service especially often.
  2. Object Detection is a part of Einstein Vision that helps marketers recognize and better classify images on social media.
  3. Einstein Forecasting is a sophisticated sales forecasting module. It provides a possibility to check whether your sales are ‘on schedule’ and go as expected, which makes it easier to monitor and manage your sales performance.

Coming soon

In the summer of 2018, Salesforce is planning to release two new modules: Einstein Prediction Builder and Einstein Bots, parts of myEinstein services.

  1. With the help of Einstein Prediction Builder, Salesforce admins will be able to create their own prediction tools based on Einstein, not only in marketing, sales or services, but all over Salesforce.
  2. And with the help of Einstein Bots, it will be possible to make a chatbot for common customer inquiries in Service Cloud. It is likely to take some load off service agents.

But that’s not all. The trick with both modules is that admins will be able to create new tools and bots with just clicks, not code.

So what?

Salesforce has made sure that their Einstein Artificial Intelligence can do lots of things: lead/opportunity/engagement scoring, email content recommendations, post classification, image recognition, blocking spam, automatic data entry, account insights, email and calendar synchronization with Salesforce, etc. And this list can really go on forever, since Salesforce is constantly coming up with new features.

But it seems that not all of Einstein’s abilities are so exceptionally AI. Some, for instance, data entry, have actually nothing to do with it. It could be done through integrations of Salesforce with, say, your website. And some, like account insights, are the results of mere (not to confuse with simple) data analytics.

Still, this is just a minor trick on Salesforce’s part, which in no way belittles the value that Einstein’s predictive capabilities can bring to your company. Especially, if tuned and tailored to your precise wants and needs.

Do you need to get an expert opinion on your Salesforce solution? Our consultants will analyze your current Salesforce capabilities and offer an optimal solution to meet your business needs.