Why Salesforce is Investing in Artificial Intelligence
Artificial Intelligence, Salesforce Best Practices

Why Salesforce is Investing in Artificial Intelligence

It’s official, the term Artificial Intelligence has dethroned Cloud Computing as the hottest buzzword of today. It seems to be attached to everything new, from computer programs, websites, apps, and even the cars we drive.

The irony is that AI or Machine Learning as it is often called is not a new concept. It was actually discovered by Alan Turing back in the 1950s, and it went through a similar period of hype/excitement in the mid-1980s as personal computers became popular.

Unfortunately, it took decades for the power of computers to advance to a point where the theory could be practically applied. You might remember the beginning of that shift as IBM’s Deep Blue beat the then world chess champion, Gary Kasparov, back in 1997. Or perhaps the more recent, well-publicised successes of both IBM’s Watson and DeepMind’s Alphago.

Either way, over the past ten years the excitement around AI has grown and as it has, it’s become harder for those of us without degrees in data science to figure out what’s real and what’s hype. What exactly can Machine Learning do for us today?

With that question in mind, and as Salesforce has just taken the extraordinary step of making their own investment in Machine Learning (TransmogrifAI) open-source, what better way to see the potential for AI than to examine why Salesforce has invested so heavily in the technology.

Machine Learning For the rest of us

Let’s start with the basics.  Machine Learning, allows computers to answer questions that we have never thought to ask them. While previous computers could tell you the square root of 2,025 (which I know is 45 because of an app on my iPhone), until recently, even the smartest computers couldn’t tell the difference between the photo of a Chihuahua with that of a muffin.

Most computer programs, from apps on your phone, to websites like Salesforce, rely on computer code written by humans. Once written the logic in the code doesn’t change much and is notoriously literal. The program can only do what a computer programmer has anticipated you asking it to do.

Machine Learning is also based on code written by humans. But unlike traditional code with its unchanging set of instructions, Machine Learning code is much more flexible. In its simplest form, it provides the computer with a specific goal, such as find Chihuahuas in photos and it then sets off looking at photos and more or less guessing if they contain a Chihuahua.

The magic of Machine Learning occurs after each guess when the program learns if it guessed correctly or not.  Regardless of success or failure, the program makes tiny adjustments to its own code in an attempt to guess better next time. It improves itself through trial and error much in the same way humans learn.

If the first guess of a well designed Machine Learning program is no better than the flip of a coin; after thousands, if not millions of guesses and tiny self-improvements, you suddenly have systems emerge that can reliably diagnose cancer from medical images or drive a car down the streets of Manhattan.

Why Should A Company Like Salesforce care about Images of Chihuahuas or muffins?

While everyone likes a cute dog picture from time to time and muffins are delicious, the application of Machine Learning goes well beyond image recognition, and Salesforce sits in a really interesting space to begin utilizing it.

Machine Learning, at least in its current form, creates value by making complex, intuitive decisions. Decisions like identifying a Chihuahua in a photo may not sound complex or intuitive, but just try and write down the specific steps you would follow to do it.

Image recognition just happens to be one of the many tasks we undertake on a daily basis that for us is rather easy and yet very complex for the typical computer program to figure out. In the world of business, we are surrounded by these sort of decisions. We have questions like:

  • Will a specific deal close?
  • Will a specific person quit?
  • How much is a company worth?
  • Will my customer pay their bills?
  • Can I afford to hire another employee?
  • What made a customer choose me over a competitor?
  • Which customers am I at risk of losing?

To answer these types of questions, companies spend millions building IT systems that gather and organize information so that humans can consume it and make the ultimate decision often based on experienced-based intuition. Salesforce is a very successful example of this sort of system as it made the consolidation of data easier and less expensive for companies of all sizes.

When deployed properly, Salesforce CRM sits at the intersection of a lot of interesting data. From prospective clients interacting with corporate websites, potential deals being pursued, customer issues and cases, contract details, and individual customer purchases all find there way into the Salesforce cloud.

It makes sense. The job of a CRM is to provide a 360-degree view of customer information so that humans can make better decisions. But with all of that data available in a single place, it’s also available for Machine Learning tools to consume and learn from as well.

Say hello to your new employee Einstein 

Enter Salesforce Einstein. In 2016 Salesforce launched Salesforce Einstein as an umbrella concept to unite all of the Machine Learning efforts it is undertaking.

While Salesforce began to use AI to power things like its search functionality in each Salesforce instance, Einstein has grown to provide scoring on new prospective customers, providing its own assessment on the chances of Opportunities closing, and with the launch of Einstein Analytics, it now provides the most powerful CRM advanced analytics tool to date.

Where is it all going?

Regardless of its potential, Machine Learning is still in its infancy. Progress and new discoveries are being made in the field on a daily basis, and that knowledge exists in the heads of too few individuals for every organization to be able to build their own Machine Learning application.

If you want to see why Salesforce is excited by the possibilities of Machine Learning, just look back at the list of intuition-based business questions. While you may never specifically ask your CRM system a question like, “Should I hire a new employee?” if you think about the data behind the decision (opportunity pipeline/forecast, historic sales, and historic employee count) you begin to see that Salesforce is probably closer than you might think to providing some interesting insights on the topic.

Hence, the real opportunity for customers of Salesforce is to rest easy knowing that regardless of what comes next, they are well positioned to benefit from Salesforce’s investments in it. While not every company has the knowledge or budget to pursue their own AI initiatives, with the cost of a simple Salesforce license your company is already embracing the possibilities of AI.

Still have Questions

If you still have questions about Machine Learning or the capabilities of Salesforce Einstein, please don’t hesitate to contact CloudMasonry. As a Salesforce consulting services provider with over a decade of hands-on experience, CloudMasonry delivers Salesforce value to organizations of all sizes.