It’s no secret to anyone that generative AI is the hot new thing in tech right now, promising to revolutionize the way humans interact with software. And, perhaps uniquely, it is a potentially transformational technology that won’t require rebuilding the infrastructure stack.
Salesforce is one a rising wave of software companies betting on the promise of generative AI. But in Salesforce’s case the bet appears to be big — and one that may finally tie together its disparate, high-profile acquisitions over the past few years in ways that could shed light on the company’s long-term vision for how its customers will make use of its full slate of offerings.
At its Salesforce World Tour NYC and Tableau Conference events earlier this month, the company unveiled a slew of new generative AI improvements and offerings for Sales Cloud, Slack, and Tableau. In collaboration with partner Accenture, Salesforce also announced a generative AI for CRM acceleration hub that will help customers scale their generative AI efforts.
“We’re thinking about how we can better utilize AI to support a sales organization in terms of creating the right strategies to approach customers,” says Juan Perez, who joined Salesforce as CIO in 2022 after more than 30 years at UPS in various technology leadership roles, including CIO and chief information and engineering officer. “How can we take that great, trusted data that we have and simplify the way in which account executives are positioning a deal in front of a customer, are positioning a solution in front of a customer?”
Gerry Murray, research director of marketing and sales technology at IDC, believes Salesforce is right on trend, noting a genuine appetite for AI in search and social services, along with enterprise cloud applications.
“Generative AI and prompt-based UIs are going to become de facto capabilities for marketers and marketing systems,” Murray says. “As the example of ChatGPT’s record-setting growth to 100 million monthly users in only two months shows, people naturally understand and prefer dialogue-based ways to interact with information.”
Generative AI is a set of algorithms that use training data to generate new content (text, images, audio, or video). ChatGPT, perhaps the most well-known generative AI application to date, is a chatbot built on OpenAI’s GPT-3.5 and GPT-4 large language models (LLMs). Like other generative pre-trained transformer (GPT) applications, it can draw on its neural network model to ask and answer questions, summarize information, and even write code.
Enterprises have been adopting the technology of late, though cautiously, including early returns on GPT-based endeavors by Unilever and CarMax, among others.
Einstein GPT everywhere
In March, Salesforce launched Einstein GPT, a generative AI CRM technology that combines Salesforce proprietary AI models with generative AI technology from partners and real-time data from the Salesforce Data Cloud. Einstein GPT ingests and unifies a company’s customer data and can then connect it to OpenAI’s models or their own model, enabling their Salesforce CRM instantiation to perform actions such as generate personalized emails for salespeople to send to customers.
Earlier this month, Salesforce added Einstein GPT for Service to help service teams generate responses to customers, automate knowledge article creation, and auto-generate case wrap-up summaries. Salesforce also unveiled Slack GPT, which enables generative AI app integrations to power capabilities such as conversation summaries and writing assistance in Slack. It also ties in with Einstein GPT to turn Slack into the conversational interface for Customer 360. The company views generative AI to be a key priority for senior IT leaders and has set aside $250 million for AI startup investment.
At Tableau Conference last week, Salesforce introduced Tableau GPT, powered by Einstein GPT, and the complementary Tableau Pulse offering. Tableau GPT changes the way users interact with Tableau, enabling them to use natural-language prompts to generate visualizations. Tableau Pulse, powered by Tableau GPT, is intended to personalize the user experience of interacting with Tableau, using AI to anticipate the data the user needs and automatically generating actionable insights.
“There are many use cases for generative AI, including creating text, image, and video brand assets, personalizing marketing at scale, and conducting conversational interactions with customers across channels,” IDC’s Murray says. “But the use case that will change the way we interact with computers is prompt-based UIs.”
Murray says users will be able to describe what they want the app or system to do, get recommendations from the AI, and create what is essentially a reusable, extensible script for a use case. Users can perform quality assurance on the script by watching the AI make the necessary menu calls and user interface clicks in a sandbox environment. Once the users are confident in how well the script works, they can put it into production.
“Prompt-based user experience is a killer app for AI because it scales and is highly observable, which will raise familiarity and trust with AI systems in general,” Murray says. “It will also end up documenting in great detail how work gets done around the enterprise and free human resources from the tedium of click-based process automation and analytics.”
The grand vision in the making?
One of the criticisms of Salesforce in recent years is that it has been slow to show how its string of multi-billion-dollar acquisitions in recent years fit together in a cohesive whole. It acquired Mulesoft for $6.5 billion in 2018, Tableau for $15.6 billion in 2019, and Slack for $27.7 billion in 2021. Some activist investors started questioning the pace of the company’s spending. The questions started taking on additional weight earlier this year when Salesforce laid off about 10% of its workforce and then disbanded its mergers and acquisitions (M&A) committee, saying it would focus on its existing business.
But these latest announcements from Salesforce begin to show how the whole might be greater than the sum of its parts, with Salesforce as a central repository for all data around a customer, Mulesoft connecting to other data sources and aiding process automation, Tableau providing insights, and Slack as the interface that surfaces the data and provides for interaction between sales teams, marketing teams, and customers.
“I’m proud to say that in the last year, almost every product that we’ve released has been a join product — you have to squint in some cases to see the Mulesoft technology,” says Brent Hayward, CEO and general manager of Mulesoft.
IDC’s Murray says that like its enterprise cloud application peers, Salesforce has found the work of building a common architecture for all its cloud applications is a heavy engineering effort.
“To be fair, the Salesforce acquisitions are still five or fewer years old and moving/merging mountains of code is a long process,” he says. “We expect Salesforce to continue its integration and interoperability efforts, especially with respect to prompt-based UIs and AI for engagement, automation, analytics, and data management because brands need to solve for continuity across the diverse array of interactions they have with customers.”
Artificial Intelligence, Enterprise Applications, Salesforce.com, Technology Industry