Amid the rapid wave of generative AI, every technology vendor was rushing to implement AI-enabled features into their offerings. Another phenomenon that came from this is the rise of copilots – chat boxes within the software’s interface that users use to instruct AI to take actions, all in natural language (i.e. the way humans naturally talk in conversation).
Watching Salesforce’s Connections event, you would have heard the phrase “copilots everywhere” come up. Salesforce is pushing the idea that multiple copilots should be working together, like an AI symphony. We’ll take a look at how this would work in practice, and why this approach will benefit your organization (as well as Salesforce’s market share).
One Copilot, Every Cloud
The image below is good to visualize where, and for what use cases, Salesforce has developed copilots – from sales and service, marketing and commerce, through to data visualization and back-end configuration.
Salesforce is pushing the idea that it’s a singular copilot (Einstein) that works across these corners of the Salesforce platform, saying “a single copilot across all of our applications…For every application across Salesforce, you’re interacting with the same visual experience, the same conversational interface”.
One would assume that each of these “clouds” wouldn’t be using different AI models. A further assumption is that your organization’s Einstein is learning from interactions and outcomes that occur in one place, and is gradually getting smarter to the context of your organization over time. By interactions, we mean users (internal/external) posing queries, and by outcomes, whether the AI output (generated text, next best action, etc.) was accepted by the user (or whether it needed further refinement).
What’s interesting to note is that Marketing Cloud is included on the slide. You may be aware that Marketing Cloud, as an acquired product, functions on its own technology stack. At Dreamforce ‘23, Salesforce unveiled its roadmap of AI-enabled features for marketing and commerce (there were a lot of them!), all while emphasizing that these features are built on Data Cloud. We’ll circle back to this later.
What Does This Mean for You?
Salesforce believes that “copilots everywhere” will result in a more natural AI adoption for you and your user base; ways of working become more fluid when natural language (conversation with AI) plays a part.
In the context of Salesforce, since the first mention of gen-AI (Einstein GPT, now copilot), they have always led with “trust” – that the data Einstein is working with never leaves the boundaries of your Salesforce org. Compare this to using something like ChatGPT, which essentially unleashes your prompt to the wider world.
The Einstein Trust Layer is intended to quell superstitions or concerns that users may have when working with gen-AI tools in their work.
Example: Marketing + Commerce Copilots
At Connections ‘24, we witnessed the power of three copilots working together (possibly for the first time). Even if you’re not interested in Marketing Cloud or Commerce Cloud, the key takeaway is how Salesforce’s copilots are no longer working in isolation – instead, they are following through end-to-end.
- Einstein Copilot for Marketers
- Einstein Copilot for Merchandisers
- Einstein Copilot for Shoppers
It all starts by asking a question. Behind the scenes, copilot builds a “plan”, compiling all the data and processes that are required to give the user an answer (a set of actions). For example, recommending the copy/timing that you send an email or an SMS to your customer.
According to Salesforce, three copilots work together in a flow, similar to the one detailed below:
- Einstein Copilot for Marketers: “An amazing digital campaign assistant”. Generate a campaign brief, in which copilot “reaches into” your customer data and brand information. If desired, this can perform other actions, such as translating content, creating email layouts and copy, and even sending emails/SMS’ out to the customer on your behalf.
- Einstein Copilot for Merchandisers: Commerce Cloud customers can converse with Einstein to set up their storefronts, write product descriptions, and optimize product listing SEO and e-commerce site conversion rates.
The “Marketers” and “Merchandisers” copilots join in harmony when looking at smarter ways to upsell/cross-sell while answering to customers’ intentions.
When bringing this intelligence into the hands of customers, that’s where Einstein Copilot for Shoppers comes in. Sitting “on top” of your e-commerce/website, browsers can interact using natural language to find the right products, delve into your catalog, and ask questions that would otherwise be gathering dust in your FAQs (frequently asked questions) section.
The vision Salesforce has showcased (with their customer, Aston Martin), is that customers will be able to interact on messaging channels (such as WhatsApp) as if they’re talking to a friend. Even with a complex product catalog, Einstein Copilot for Shoppers can surface recommended product add-ons and determine whether modifications can be made to the order based on production and shipping status data.
Where Data Cloud Comes In
We’ve all been hearing plenty about Data Cloud. As mentioned earlier, Data Cloud underpins the various copilot experiences.
According to Salesforce’s recent report, only 31% of marketers are fully satisfied with their ability to unify their customer data across sources. Data Cloud, handling data harmonization, ensures that there’s a “common view” of an individual customer – in other words, connecting everything “under the covers”.
Salesforce used the term “trapped data” to describe sources that are not accessible in a traditional CRM platform (unless ported over via integrations). Salesforce has forged partnerships with key technology players in order to achieve “zero-copy architecture”, where data in sources external to the Salesforce platform can be referenced, but not actually stored on the platform.
This avoids data center processing (a positive for the environment), and storage costs. The takeaway here is that this additional data makes for a more powerful copilot experience, wherever you’re interfacing with Einstein.
Summary
As demonstrated with the marketing and commerce copilots seamlessly working together, Salesforce is pushing the idea that it’s all one copilot (Einstein) that works across all corners of the Salesforce platform.
We can now see that the case for Data Cloud is getting stronger. Having invested in developing Data Cloud into what it is today, Salesforce’s obsession with Data Cloud is getting clearer – powering these tools and leading the AI revolution.