An interview with Rahul Auradkar, Salesforce executive vice president and general manager of unified data services and Einstein.

Deborah Yao, Editor

January 22, 2024

9 Min Read
Photo of a laptop with letters CRM in text
Getty Images

Rahul Auradkar, executive vice president and general manager of unified data services and Einstein at Salesforce, joins the AI Business podcast to discuss the Einstein platform and other key AI initiatives.

Listen to the podcast below or read the edited transcript.

Can you tell us a little bit about what you do at Salesforce?

I'm responsible for Data Cloud, one of our fastest growing organic innovations in the history of the company. It delivers on our Customer 360 vision with AI, data, CRM and trust. The second thing I'm responsible for is Einstein, our AI platform, and my charter includes the buildout and growth of the product by partnering with my engineering counterparts and business growth by partnering with sales, marketing, customer success, etc.

At a high level, can you share what Einstein is, as well as tell us about your investments and partnerships in AI? Particularly generative AI?

Our approach of AI plus data plus CRM, which is a virtuous circle, and one feeds the other. All of this is grounded in trust. In order for us to deliver the Customer 360 vision, what we have been doing as a company is building on our successful predictive AI investments using Einstein and we have been the first to introduce Gen AI-driven CRM products to the market.

More than 80% of IT leaders in our survey are telling us that generative AI will have a prominent role in their organizations. If you think about AI itself, the AI revolution is a Gen AI revolution, it is a data revolution. AI is only as good as its data strategy. To that end, we have been investing heavily in Data Cloud, which provides a single view of customers, and it powers Einstein and also generative AI and predictive AI.

The second broad area of investment we have is on trust. The AI revolution is also a trust revolution. It is essentially our foundation in which we deliver predictive and generative AI for our customers. We are also innovating heavily across what we refer to as clouds, our CRM applications, and whether it is on the sales or service etc.

Why use AI in CRM? What problem are you solving?

This is a customer revolution. As an example, AI transforms the way customers interact with brands, the way we innovate. Let's take the example of PenFed (Pentagon Federal Credit Union). PenFed recently invested in Salesforce Einstein’s generative AI capabilities, which it plans to use as an as an assistant to provide responses to chats or member questions, beginning with internal employees and then expanding it out to all its members. … (And) we are doing that in a trusted way. We are doing it in a manner in which it provides outsize value for our customers so they can deliver delightful experiences for their customers. And we are seeing some outstanding results with respect to productivity and efficiency and value creation using generative AI.

Tell me about your generative AI efforts. What is your goal?

The purpose and goal really starts fundamentally with our customers’ need for productivity and the ability to use the data at a speed to deliver CRM and downstream value to our customers. From a predictive and generative standpoint, the goal is really to give our customers a way to deliver predictions and generations of experiences downstream with their customers for personalization at a very large, massive scale, and doing it in a very trusted manner.

Even though more than 80% of our customers are indicating that they are going to be using generative AI in some shape or form to deliver value to their customers downstream, only about 30% are getting started right now for various different reasons, including the issue of trust and the fast moving pace at which this technology is moving. Our goal is to make it easier for customers to use generative AI to deliver those products and its benefits through a combination of Data Cloud and our trusted Einstein platform.

Some customers may be hesitant about using generative AI because there are some very well-known risks associated with it. How do you address that?

In our survey, over 60% of the people who want to use generative AI, they are more concerned about safety, security and trust. This is where we are differentiated. Salesforce AI is grounded in trust. As a company, our core values are trust, customer success, innovation, sustainability and equality. All of that grounds our AI strategy, the Einstein layer that we have built is deeply integrated into our architecture. Essentially, our customers can trust us to keep their customers’ data safe.

Secondly, there is ethical use. Salesforce has taken a leadership role in the ethical use of AI in our products and services. In fact, we have an office of ethical use of AI to drive our AI innovation with trust guidelines and capabilities that protect not just our customers, but also their customers’ data as well.

Can you be more specific about how you safeguard against these generative AI risks? What things do you put in place?

Let's talk about the Einstein trust layer. The secure AI architecture is natively built into the Salesforce platform. It is designed for enterprise security standards, the trust layer lets the team step into generative AI without compromising customer data. The trust layer essentially has a few things. First, it grounds and enriches generative prompts through an integration into Salesforce Data Cloud. And within Data Cloud, we have all sorts of protections for masking, etc. We have also audit trails. The audit trails allow us to make the platform compliant and provides visibility and control on how data is being used, and also has the ability for our customers and auditors to take a look at how data is getting used for their AI.

We also have agreements with LLMs for what we refer to as zero retention. That means that the LLM providers are not retaining any data to train their models using our customers’ data. The next thing is toxicity when it comes to generative AI and LLM. That can hurt brands with respect to responses that generative AI is providing. So we have toxicity detection: the ability for us to mute toxic responses from generative AI. We are also doing masking, for example, with PII or personally identifiable information, we do not send it out to external LLMs that are generating the responses. Those are some of the areas that are deeply integrated into Data Cloud, deeply integration into our AI platform for using Einstein.

Can you talk more about Einstein Copilot and how it enables everyone to be data scientists?

It is a conversational assistant for Salesforce applications. In essence, it is baked into the flow of work − the idea that you have the ability to have no-code type AI within the flow of work. It can be used for embedding AI assistants into websites, to power chat, messaging platforms like Slack, WhatsApp, SMS, etc. Copilot allows you to interact in a natural language, which then uses AI to generate trusted and accurate recommendations, guidance and content as well. It is not just about drafting emails; it also can do a variety of different things. We are excited about it, it's going to get it's going to be ready in the first quarter of (2024).

Can you give us a sense of the reception of the announcements made at (the Salesforce event) Dreamforce from your clients?

It has just been outsized. Customers are telling us that they want to use generative AI to deliver some level of value to the downstream customers. … We are being very thoughtful and deliberate as we take our customers through this journey, but the reception has just been outsized. And it is not just the customers, but also analysts and the press. The reception has been completely well beyond our imagination of positivity.

Can you give me an example in more detail about how this autonomous CRM would work?

You can think about an autonomous CRM as touching the customer through multiple different touchpoints and channels and it is seamless to the customer − in terms of responses the customer gets. Once you are done with that, it turns into a sales lead and all of that is personalized inside of the system. The interaction between the salesperson and the lead happens automatically through either different modalities, whether it is video or audio, or different modes, whether it is with the cell phone or browser, and the buying experience also gets automated to the point where the system knows quite well about what you are trying to buy.

And it does predictions along the way it and those predictions are highly personalized. Subsequently, once you are part of the company's customer base, then all of the interactions that you have with respect to service are also automated, whether it is case deflection when you are having an issue or whether it is related to an automated response or a recommended response for an agent to interact with you in a very personalized way or automation. Automating a task, for example, billing tasks, or whether it is for a refund or return.

There is a lot of talk about how AI could be disruptive, but I think it is one of the most constructive things that is going to happen in our industry. When it comes to CRM, providing relevant customer experiences that are personalized through multiple different channels, multiple different touch points, and providing seamless experiences, all of the things that we have been talking about get accelerated in a very fast timeframe. The amount of value we are creating is so much higher than the disruption. So I'm highly optimistic about what AI can do generally, but more importantly, highly optimistic for what it can do for CRM.

The second one is Google, Amazon, and other big tech companies are all coming together for the betterment of the customer. When it comes to data and AI, the boundaries are getting erased and that's the most exciting thing. This has dramatically shifted to the point where now, data sitting in Google, for example, today can seamlessly show up through our innovation (bring your own data lake into Salesforce), which we refer to as cross joints or the ability to have one big query data joining with Salesforce data which in turn joins up with Google data. To me that is the most exciting because for the longest time we have been talking about companies working together to deliver value to customers. This is getting accelerated at a pace which we have never seen before.

Read more about:

ChatGPT / Generative AI

About the Author(s)

Deborah Yao


Deborah Yao runs the day-to-day operations of AI Business. She is a Stanford grad who has worked at Amazon, Wharton School and Associated Press.

Keep up with the ever-evolving AI landscape
Unlock exclusive AI content by subscribing to our newsletter!!

You May Also Like