Since Genie, now Data Cloud, was announced at Dreamforce ‘22, Salesforce has gone on a frenzy to place this product at the forefront of their mission.
We used to talk about Customer 360, that is, integrating all of your data into your CRM to create a 360-degree view of your customer. But now the tact has changed, Salesforce talk about AI + CRM + Data in order to succeed in 2024 and beyond – but what does this actually mean? Let’s explore…
Data Cloud: The Fastest Growing Product Ever
Salesforce announced their Q4 earnings towards the end of last month, beating analyst expectations by growing 10.8% YoY, but also forecasting single-digit growth of 8.6%, which will probably be their lowest yearly growth on record. Despite this, their margins are increasing, they are paying a dividend for the first time, and Wall Street is happy; Salesforce’s stock is up 70% over the past year.
Salesforce has been boasting about Data Cloud for a while, calling it “the greatest Salesforce innovation in the company’s history” at Dreamforce ’22, but now, we have some cold hard facts on the success of the product.
In the Q4 earnings call at the end of February, Marc Benioff revealed some interesting insights about their golden child in the last quarter…
- Data Cloud is approaching $400M in ARR, growing at nearly 90% year over year.
- It’s the fastest-growing organic product in the history of Salesforce.
- In Q4, 25% of deals over $1M included Data Cloud.
- Salesforce have recently added 1000 net new customers.
- 7 trillion records were ingested into Data Cloud.
Salesforce also had something else to shout about in the earnings call: the fact that Gartner had placed them as a clear favorite in their Magic Quadrant, beating out competitors like Twilio and Adobe. This is even more impressive considering the fact Salesforce only received their CDP certification from the CDP Institute in the summer of last year, which certifies CDP platforms have a certain level of expected functionality.
What is a CDP & Why Should I Care?
Unless you are a marketer or have been involved in a CDP project, you probably don’t have a clue what Salesforce is talking about with their Data Cloud, and this is understandable. After all, we’ve got a CRM, so why do we need yet another system to implement and spend money on?
Well, that answer lies in the reinvigorated conversation of data. Whilst Data has always been important, we are collecting more than ever. CRMs are unable to support the needs and requirements of modern data collection, storage, and dissemination.
CRMs are primarily for sales, customer support, and marketing teams, recording customer details, and information about their interactions such as sales calls, or support ticket requests. But there is vastly more information than this collected within a company. We have in-product user behavior, offline interactions, payments, products purchased, website pages visited, etc…
A CDP, or Customer Data Platform, is designed to store customer data across every touchpoint during the entire customer lifecycle. This means that instead of the CRM being the single source of truth, the CDP is considered to have this view, as more data is being ingested from various sources.
Salesforce themselves have tried to make the CRM as the source of truth for many years, placing Sales and Service Cloud at the heart of customers’ businesses, as well as using products such as MuleSoft to integrate silos of data, Marketing Cloud to bring in customer interactions, and Commerce Cloud to pull in customer purchases.
But Salesforce have said in a recent Q&A about Data Cloud that…
”Despite efforts to centralize data and build a 360-degree customer view in their CRM, most company data remains disconnected, and therefore unusable or only partially effective to improve the customer experience”.
In my opinion, two foundational aspects make CDPs and Data Cloud particularly special…
Firstly, it’s their ability to easily ingest data from multiple sources. Salesforce have partnered with Snowflake, Databricks, AWS, and Google to enable zero-copy/ETL with Data Cloud. This means that data is accessible and viewable in Data Cloud, but is not actually moved, which supports data integrity.
Salesforce also has connectors to the vast majority of their products, as well as third-party products such as Amazon S3. In addition, MuleSoft can be used for more unique or custom databases. Data Cloud and CDPs also support unstructured data, that is, information of meeting notes, phone calls, and PDFs.
This is different from Sales and Service Cloud, which provides some out-of-the-box connectors, but not to the same degree.
Secondly, Data Cloud gives customers the ability to harmonize the data, mapping objects and fields to a common model, and then using Identify Resolution to create a single view of the customer – the ultimate goal of many businesses for decades.
But there is one element we are yet to chat about, and that is artificial intelligence in the form of Einstein 1, the new name for the Salesforce Customer 360 platform.
How Do Einstein 1 & AI Fit into Data Cloud?
Salesforce Einstein was first announced at Dreamforce 2016, and included several predictive analysis tools that told you how hot a lead was, or how likely an Opportunity was to close. Although this was very exciting at the time, the excitement quickly died away, and I don’t think the products were as successful as Salesforce would have hoped.
Over the years leading up to 2016, Salesforce spent billions of dollars on acquisitions, which included $75M spent on AI start-ups, and $700M spent on Krux, the product that would eventually become Data Cloud.
I’ve always thought that Salesforce are a bit ahead of the curve, and I mean you have to be, right? They pioneered cloud computing back in 1999, and have been innovating and acquiring ever since to become the leading company for CRM, sales, service, and marketing. Salesforce and Benioff may have seen the future as far back as 2016, acquiring companies that would eventually have an impact on their 2024 strategy, Einstein 1.
The general message behind Salesforce’s AI + CRM + Data messaging is that for AI to work in the world of B2B, you need data collected automatically across the entire company in the form of a CDP, and you, of course, need your CRM for your customer-facing teams to work out of, and utilize the AI insights.
Specialist vs. Generalist AI
It seems like the whole world is talking about specialist vs. generalists at the moment, including in my most recent analysis of the Salesforce job market.
But I think the conversation in regard to AI is extremely important. Many analysts agree that OpenAI’s ChatGPT, whilst impressive, provides the most generic answers possible to any question. This isn’t surprising considering ChatGPT’s LLM is built using information across the entire internet and is not built for any specific purpose.
This is why, initially, I don’t think the generative aspect of AI is going to be that useful for businesses in phase one of this transformation.
I always like to relate this conversation back to using ChatGPT to generate a sales email. Who exactly should be using this functionality within a business? Senior salespeople (I would hope) would be able to craft a much more impressive sales email than using a generic LLM. So does this mean junior sales users should use it?
In my opinion, they won’t be able to properly evaluate if the very generic sales email is good or not, and in this learning stage of their career, probably won’t teach them much.
The truly impressive part of artificial intelligence will surface itself when large amounts of data, are combined with an LLM designed for a specific purpose. This is exactly how Salesforce have positioned their AI tech stack…
The Power of Einstein 1
Einstein 1 has a mixture of artificial intelligence tools on the platform, and it kind of mirrors the way the core Salesforce platform is built, standardized and custom.
We have out of the box AI features such as sales email generation in Sales Cloud, and service replies in Service Cloud. But then we have tools such as Copilot, Prompt, and Model Builder, three tools contained in the “Einstein 1 Studio”, that allow us to build custom AI functionality and roll it out to our users, which can also be integrated with other Salesforce functionality.
For example, we could create a Prompt on an Opportunity that enables an Account Executive to summarize similar closed won deals to overcome a complex objection, also finding out what kinds of products and discounts were used on those deals.
However, whilst this information could be useful, as we’ve established, it depends on the Large Language Model you are using which will determine how specific or generic the answer to a question is. Well, Salesforce have an answer to that too.
Out of the box, Einstein 1 Studio will allow you to use Salesforce’s own proprietary LLM which is presumably the latest “XGen-7B”. Salesforce has been somewhat secretly building LLMs for over a decade, and their latest XGen-7B allows you to input more data than traditional LLMs, as well as receive longer outputs.
Finally, if you are interested in using other LLMs available on the market, or even using one that your own company has created using proprietary data, then Salesforce gives you the ability to switch, using their Model Builder. This gives Salesforce a completely agnostic approach to LLMs.
Summary
Salesforce often takes risky bets, throwing their resources into the ‘flavor of the month’ such as Blockchain and NFTs. But AI doesn’t seem like a trend that will wither and die. It’s probably only a matter of time before it starts to enable every aspect of our personal and business lives.
If everything goes well, Salesforce’s big bet on Data Cloud and Einstein could provide a huge second wind to the company’s growth story. Companies know that they need to throw money at some kind of AI product to stay relevant and outperform competitors, even if the space is still so green at the moment.
In my opinion, the true power of Salesforce AI products will come from the Einstein 1 Studio, which has only just been released. It will be exciting to see what the Salesforce ecosystem and create, and all the wild and wonderful tools that will be created using this selection of tools.