Introduction to Artificial Intelligence
Artificial intelligence, or AI, is the rage today. Generative AI, which can ‘feel like’ the machines are understanding us and talking back, has been the cause of the outburst of public awareness and the light-speed growth in the AI meme. However, what is now called ‘AI’ was built on other forms of machine learning that are now considered part of this overarching category. This is because the software industry latches on to the latest buzzword in order to market itself.
The reality is that AI has been evolving for more than 50 years. The basic elements of neural networks, that are at the heart of today’s AI, were first invented in 1958 by Frank Rosenblatt of Cornell Aeronautical Laboratory. Since then, there have been four epochs in the development of AI, with large gaps of apparent inactivity in between until some new technological breakthrough occurred. But AI, as we understand the innovations of today’s fourth phase, really began to evolve in commercial markets in the early 2000’s (Figure 1).
It started with something called statistical learning, and became prevalent as the volume of data in online services from Google, Amazon, Facebook, Yahoo!, and others exploded into petabytes. This provided enough data to begin to make predictions at scale. Statistical learning turned into machine learning in the late aughts (the decade between 2000 and 2009), which turned into deep learning around 2014, which subsequently turned into reinforcement learning around 2018, and became what we now call AI with the evolution of generative AI in 2022/2023.
Predictive Machine Learning
So today everything is called AI, but the reality is that a lot of what we call AI today is what we called machine learning only 10 years ago. It can make data predictions in one of two ways. Supervised learning has a target you can test a model against. For example, an auto manufacturer has a history of auto repairs in new cars we build over time. We want to predict the likelihood of an electrical problem with our new auto line. We use the old data as the target and then try to build a machine learning model to predict the actual historical data. If we are accurate enough, we then apply that machine learning model to the new car line/year and its electrical components and predict the likelihood of it having electrical failures. If the model is accurate and the underlying components haven’t changed drastically, our prediction should be pretty accurate.
With unsupervised learning, you don’t have target data to train the model. Instead, data scientists have found ways to use “clustering” techniques to identify and classify patterns in the data. The algorithm underlying the model looks for these patterns and puts a data point in one of the clusters based on how it “ranks” on the combination of features that are considered when the model is built.
Salesforce and AI
This history is important because it gives context to Salesforce’s investment in and deployment of AI into its platform. The fact is Salesforce’s commitment to what we call AI today began in 2014. That is when Marc Benioff (now) famously stated at a company meeting that “Salesforce will become an AI-first company.” Even though AI as we understand it today was not generally called that or in broad commercial use at that time, Benioff, like a few of the top visionary leaders in the technology industry, saw that it would be there in a timeframe that could make it the “North Star” to which company mid- to long-term strategy would point.
So Salesforce set off on an AI-focused journey. But it started with typical machine learning capabilities, and evolved them slowly to the point today that there is a large component of the Salesforce platform(s) that is truly-AI driven. But in reality, the suite of capabilities in Salesforce is a mixture of machine learning, deep learning, reinforcement learning, and AI techniques.
While the historical evolution of the technology is one reason, the more important reason is that Salesforce in its AI suite uses the best tool to solve the right job. Sometimes the best tools use basic machine learning. For example, Saleforce’s B2C Commerce Einstein API has a recommendation engine built in. But as the API documentation states “Einstein Product Recommendations uses machine learning to power 1:1 personalized shopping experiences across every customer touchpoint.” Not Salesforce AI. Salesforce machine learning.
Why? Because recommendation engine technology based on machine learning techniques is quite evolved and reasonably accurate and because this API has been around since 2017 when machine learning-based recommenders were state-of-the-art. Changing it right now is probably not a high priority. Can generative AI algorithms potentially improve product recommendations? No doubt, but that requires a great deal of time and testing. As the old adage goes, “If it ain’t broke, don’t fix it.” As a result, as we talk about Salesforce AI, we will use it as shorthand for all the data science-based technologies underlying Salesforce’s platform. When we want to talk about specific capabilities, we will refer to them based on the particular data science approach their algorithms are based on (if that information is publicly available).
The History of Salesforce AI Capabilities
The table below provides a good sense of how Salesforce AI’s capabilities have evolved and how the Salesforce Einstein Platform became its core.
| Date | Type of Action | Name of Item | Description |
|---|---|---|---|
| 2014 | Product Release | Salesforce Wave Analytics | Introduced business intelligence and data visualization capabilities. This laid the foundation for future SalesForce AI features by providing insights and data exploration tools. |
| 2015 | Acquisition | RelateIQ | Acquired RelateIQ, a company specializing in relationship intelligence. This provided Salesforce AI with capabilities like identifying key contacts and predicting next steps in customer interactions. |
| 2016 | Acquisition | Demodata | Acquired Demodata, a social listening and customer sentiment analysis platform. This allowed Salesforce to integrate social media data into its AI models and gain a deeper understanding of customer sentiment. |
| 2017 | Product Release | Einstein | Launched the now unified AI platform for Salesforce, including features like predictive lead scoring, opportunity insights, and automated workflows. This marked a significant step towards offering comprehensive AI functionalities within Salesforce. |
| 2018 | Acquisition | BeyondCore | Acquired BeyondCore, a company with expertise in natural language processing (NLP). This enhanced Einstein’s capabilities in areas like sentiment analysis and text summarization, allowing for deeper analysis of customer data. |
| 2019 | Product Release | Einstein Prediction Builder | Introduced a tool to Salesforce AI that enabled users to create custom AI models without coding. This democratized access to AI functionalities within Salesforce for less technical users. |
| 2021 | Acquisition | Tableau | Acquired Tableau, a leader in advanced data visualization and business intelligence. This further strengthened Salesforce AI’s ability to analyze data and integrate insights into AI-powered features. |
| 2023 | Product Release | Generative AI Integration | Began integrating generative AI capabilities into the Salesforce AI platform. This allows for functionalities like automated content generation and conversation summarization, potentially transforming customer interactions. |
The Philosophy Behind Salesforce AI
As leaders in leveraging AI, DataGroomr takes a conscious approach to its use. AI, for all its benefits, has many potential downsides. Misuse of private data, intentional or unintentional biases in the algorithms, and the potential for generating just plain wrong results without an understanding of why are only a few. Anyone who is doing serious work with AI has a set of underlying principles they follow to try to ensure that these challenges are handled appropriately. Salesforce has been especially conscientious about this (as have we) and has a five-point set of AI principles that it follows (Figure 2). It began work on these in 2018 and spent almost a year getting feedback from both within and outside the company before publishing.
Source: https://blog.salesforceairesearch.com/meet-salesforces-trusted-ai-principles/
- Responsible. Salesforce will ensure that its application of AI uses data appropriately, safeguards private data, and ensures that the use of AI does not impose on the basic rights of individuals or groups. Although this might be too strong a statement, the concept of responsibility behind Salesforce AI is their equivalent of doctors’ Hippocratic Oath: “first do no harm.”
- Accountable. The teams building Salesforce AI’s capabilities will be accountable to the company, its customers, its partners, and other audiences which have a vested interest in Saleforce’s creation and use of AI capabilities. They will create mechanisms, both manual and automated, to solicit continuous feedback on the performance and impact of their work.
- Transparent. This has two aspects. The more common usage of transparency around AI has to do with being able to understand why an algorithm came out with an outcome. What was the mechanic that caused any bias that a Salesforce AI team was seeing in a result? But this principle also extends to cover the application of these tools by end users. In this case for Salesforce AI, transparency means providing a clear means to understand what a specific AI-driven API or capability does and how to implement/employ it.
- Empowering. Salesforce AI’s capabilities have a key imperative: help Salesforce’s customers and their clients succeed. That success can come in the form of growing revenue, providing more jobs, or improving the lives of people in communities generally.
- Inclusive. The term ‘inclusive’ when used in AI is often referring to unintentional biases in algorithms that tend to favor one gender, ethnic group, age group, or demographic over others. For example, an algorithm that determines who is qualified for a position should not tend to favor men over women or Hispanics over Asians.
But as research has shown time-and-time again, this kind of bias can happen in algorithmic development quite consistently. These unintentional biases have given rise to a whole market dedicated to AI governance tools to identify and remove such biases.
Salesforce’s goal is to have this form of governance built-in to Salesforce AI from the get-go.
The Components of Salesforce AI
Trying to get a holistic view of Salesforce AI can be a bit daunting. Because the functionality has grown over time, there are different descriptions of the capabilities from different times or different perspectives, depending on which group within Salesforce is providing the content for which audience. Figure 3 gives our take on the Salesforce AI platform(s), now all contained within the Einstein 1 brand. We have purposely tried to use slightly (but not dramatically) different terminology in this diagram because we are trying to express a more holistic view to this architecture. For example, nowhere in their cloud architecture does Salesforce call out APIs for developers that leverage aspects of Einstein, but they are very clearly available to developers in Salesforce’s developer portal.
We will use visuals from Salesforce in later articles to drill into the capabilities of Salesforce AI as they shed light not only on Salesforce AI’s specific capabilities but also on the underlying services Salesforce Einstein draws upon (e.g. real-time data processing) to deliver results. But for the purposes of this introduction, we will use Figure 3 to provide an overview.
We will leave it here for now. In our next two posts, we will cover the top layers containing the Salesforce AI Cloud and Salesforce ML- and AI-Enabled APIs. The other elements of Figure 3 will be covered in subsequent posts.
