Earlier this week, Salesforce CEO and Co-Founder Marc Benioff hosted the Salesforce AI Day, with the aim to unveil the future of trusted enterprise AI, alongside live demos and insightful conversations.
The main announcement, Salesforce AI Cloud, is claimed to be the fastest and most trusted way for Salesforce customers to deliver generative experiences across all their applications and workflows. Also introduced was the Einstein GPT Trust Layer, which aims to resolve concerns over adopting generative AI.
After reiterating Salesforce’s backstory (having pioneered AI in CRM since 2014), and sharing their impressive trajectory with the LLMs (large-language models) that will power Einstein GPT, we were given a sneak peek at some of the tools we can get our hands on in the next 2-3 months – yes, 16 new releases, no less!
To achieve this blast of releases in quick succession, Salesforce have mobilized the entire company to work towards shipping these AI enabled products. Now, their hard work is paying off. Here’s what’s coming up in the “Summer of AI”.
Sales Email Generation With Page Context
Sales GPT | In pilot June ‘23; Generally available July ‘23
Auto-generates email copy using recommended email contents, not just based on what reps have done in the past, but also what led to successful outcomes. For example, it doesn’t just select popular phrases from “sent” emails, but analyzes whether Opportunities eventually became “Closed Won”.
This allows the rep to ask Einstein to generate an email that’s personalized and relevant – all by using natural language prompts (i.e. the way we converse, naturally). This can be copied straight into the email builder or adapted to match the tone of the email, such as making the copy sound less formal.
Service GPT | Generally available June ‘23
Auto-generate personalized agent chat replies within the Service Console. Agents then have the choice to either send the reply immediately or edit the message.
The Einstein Replies widget also weaves in next best actions, such as “add to cart”, to take further legwork out of a service agent’s day.
Domain-Grounded Reply Recommendations
Service GPT | In pilot July ‘23
Arriving one month later is Reply Recommendations that are “Domain-Grounded”. Salesforce has designed AI Cloud (which is powered by Einstein GPT) so customers are reassured that they can use generative AI in a trusted, safe, and secure manner.
The definition of “grounded” in an AI context means that the results can be verified by authoritative sources:
This is when we could see Reply Recommendations get even more accurate. Should the customer ask another question, for example, with the specifics of where they are going to use a garment, GPT generates another ready-to-use response, including information about suitability for the location the customer is heading to (by tapping into external climate data).
Service GPT | Generally available June ‘23
During case wrap-up, a summary is automatically generated of the conversation with the customer, so the agent doesn’t need to spend unnecessary time documenting it. An action will appear to swiftly update the Case summary.
Knowledge Email Creation & Knowledge Creation v2
Service GPT | In pilot June ‘23 (generally available August ‘23) and July ‘23, respectively.
Einstein asks the agent whether they would like to create a knowledge article based on the chat conversation, meaning that this information comes out of the case comments, and can be opened up to other agents and customers.
Subject Line Generation
Marketing GPT | In pilot June ‘23
Not only can you generate messaging faster, but the end result reflects what has done well in your organization in past campaigns. This content delivery is powered by Data Cloud.
Adjust copy and/or refine the tone of subject lines rapidly. For example, generate three subject line variations, based on top performing emails. The marketer has the decision making power to accept the suggestions, or they can opt to A/B test.
This will also be applicable to landing page generation in the future:
Generative Search Answers for Service Agents
Service GPT | In pilot June ‘23
Just like Reply Recommendations, our hunch is that this could be another widget in the Service Console that will enable service agents to get answers to customer questions, right at their fingertips. Again, this is likely that agents can receive answers by asking prompts in natural language (i.e. the way we converse, naturally).
Einstein GPT for Developers v1
Platform (Code Builder) | In pilot June ‘23
Einstein GPT will improve developer productivity with an AI chat assistant to generate code and ask questions for languages like Apex. This is powered by Salesforce’s proprietary LLM (large-language model), which shows up in the generated code as “This code was generated by Einstein GPT using Salesforce LLM.”
Service GPT | In pilot July ‘23
With minimal information on this particular feature, one could assume that this will have similar capabilities to the Case summarization mentioned before, but applicable to Task records. Stay tuned for more information!
Commerce GPT | Generally available July ‘23
By accelerating building multiple versions of descriptions and translations, new storefronts are faster than ever to ‘stand up’. It can also automatically fill in missing catalog data for merchants and revolutionize the customer experience with auto-generated product descriptions tailored to every buyer.
Data Cloud | In pilot July ‘23
Marketers are given the ability to create audience segments quickly and improve targeting using natural language prompts to query Data Cloud, Salesforce’s CDP offering.
Also, AI-driven recommendations guide marketers to generate the most impactful segments that are predicted to increase overall campaign engagement.
This accelerates a significant chunk of what has been time-consuming with regards to segmentation – namely discovering what has worked historically by scouring through rows of data to perform analysis, and writing queries for audience segmentation.
Tableau GPT | In pilot August ‘23
Tableau Pulse is a reimagining of Tableau, powered by Tableau GPT – an AI-powered, personalized experience that transforms how people work with data.
Surface insights in both natural language and visual format. You can either take guidance from Tableau GPT prompts or ask your own questions.
AI Cloud Roadmap (June-August ‘23) by Cloud
|Feature||Cloud||In Pilot||Generally Available|
|Sales Email Generation with Page Context||Sales GPT||June ‘23||July ‘23|
|Reply Recommendations||Service GPT||–||June ‘23|
|Domain-Grounded Reply Recommendations||Service GPT||July ‘23||TBC|
|Summarization||Service GPT||–||June ‘23|
|Knowledge Email Creation||Service GPT||June ‘23||August ‘23|
|Knowledge Creation v2||Service GPT||July ‘23||TBC|
|Generative Search Answers for Service Agents||Service GPT||June ‘23||TBC|
|Task Summaries||Service GPT||July ‘23||TBC|
|Einstein GPT for Developers v1||Code Builder||June ‘23||TBC|
|Subject Line Generation||Marketing GPT||June ‘23||TBC|
|Product Descriptions||Commerce GPT||July ‘23||TBC|
|Segment Generation||Data Cloud||July ‘23||TBC|
|Insights Summaries||Tableau GPT||August ‘23||TBC|
When it comes to the Salesforce platform, every major announcement this year has included some element of “GPT”. First came Einstein GPT (the umbrella brand for GPT technology on the Salesforce platform), with the initial announcement featuring Einstein GPT for Sales, Einstein GPT for Service, and Einstein GPT for Marketing.
Then we saw Slack GPT, which was followed by Tableau GPT. Most recently, Marketing GPT and Commerce GPT have joined the group, each with their own sets of use cases for marketers, merchandisers, and shoppers.
Within each of these GPT-enabled products, there’s an array of individual use cases ideal for the user base they are targeting. Now, you know which of the 16 new tools we can get our hands on in the next 2-3 months.