Salesforce’s business intelligence platform, Tableau, is getting generative AI capabilities in the form of Tableau GPT, which is built on the company’s own Einstein GPT AI engine, which is also integrated into other products like Slack.
“Tableau GPT can power and automate the analysis, exploration, sharing, consumption and more of data. Generative AI engines, for example, make analyzing data feel like a conversation through a chatbot, rather than drag and drop.” , brings a number of very exciting use cases.”
“Other use cases include predicting what questions a user might ask based on what the engine already has in the data, or gaining hundreds of insights and using highly understandable summaries.” and describe them,” says Arellano.
Einstein GPT, the foundation of Tableau GPT, is made up of a variety of large-scale language models (LLMs), including OpenAI, Cohere, and internal proprietary Salesforce models, says Sanjeev, principal analyst at independent consulting firm SanjMo. Mohan said.
According to Hyoun Park, principal analyst at Amalgam Insights, these internal models were driven by Salesforce’s investment in companies with natural language processing capabilities and insights into how companies do data analysis.
“Tableau previously acquired Narrative Science, a natural language generation solution for analytics. Additionally, Salesforce has made strong investments in data science over the years, including BeyondCore, Metamind, and Datorama. We also have hundreds of data scientists in-house,” Park said.
Additionally, Arellano said Tableau GPT is given a data security and governance layer to protect corporate data from internal and external data breaches and unauthorized access.
Governance and security additions may stem from Salesforce’s efforts to build trust among customers, especially at a time when companies are banning the use of OpenAI’s ChatGPT over data leak concerns, analysts said. the list said.
“These layers protect users who fear that their prompts will be used to retrain their LLMs. They can also prevent LLM hallucinations,” said Mohan of SanjMo.
Tableau GPT will be available in a pilot later this year, according to the company.
Proactive data analysis with Tableau Pulse
Salesforce also released a new flavor of data analytics under a service called Tableau Pulse. This provides a proactive analysis, the company says.
“It’s like a personal guide for your data that knows your data. It recognizes the goals you’re trying to achieve with your data. It helps you do that,” Arellano said.
Tableau Pulse will also use Tableau GPT to help enterprise users make better, faster decisions using automated analysis of personalized metrics “in a way that makes sense.” .
Use cases include alerts when there are anomalous changes in data or metrics, helping users drill down into the reasons for anomalies, the company said.
These insights can be further shared with colleagues via collaboration platforms such as Jira and Slack to find solutions, Salesforce added.
David Menninger, research director at Ventana Research, said:
However, analysts say Tableau may be trying to catch up with other vendors.
Doug Henschen, Principal Analyst at Constellation Research, said:
According to Hengsheng, cloud-based products tend to have an advantage when it comes to analyzing user behavior and data interactions at scale.
“Products that started as server-based products like Tableau typically took time to develop charts and personalization features that could be delivered consistently across both cloud and on-premises deployments,” Henschen said. says.
Many vendors offer automated insights, but adding narratives generated by generative AI “makes these insights more complete and easier to deliver in multiple languages.” said Ventana’s Menninger.
The company says Tableau Pulse will be available in a pilot later this year.
Data Cloud for Tableau to unify data for analysis
In addition to Tableau Pulse, Salesforce offers Data Cloud for Tableau to consolidate and analyze enterprise data.
At Dreamforce last September, Genie.
“Tableau allows us to visualize all of a company’s customer data so that users can more easily explore and find insights. Data Cloud also supports zero-copy data sharing, which means Users will be able to virtualize their Data Cloud data in other databases and make it readily available to everyone,” the company said in a statement.
The Data Cloud for Tableau will also include data query capabilities, the company added.
According to Henschen, there are many business benefits that Data Cloud for Tableau can provide.
“Benefits include bringing all heterogeneous data together, decoupling compute and storage decisions, and enabling different kinds of analytics and different use cases to the data cloud without replication or redundant copies of data. ,” says Henschen.
According to Park, Salesforce’s attempt to integrate its Data Cloud with Tableau may be due to Tableau’s core analytical discovery capabilities reaching their limits.
“There is increasing pressure to support larger-scale analytical use cases that push data management and data warehousing. We want to be the source of the master data that we access,” says Park.
However, a data cloud for Tableau is part of a strategy to compete with data lakehouse and data warehouse vendors and an effort to own or control more data, Menninger said. says. The integration of Tableau and Data Cloud will put the company in direct competition with the likes of Qlik, Tibco IBM, Oracle and SAP, analysts said.
Data Cloud for Tableau will be available later this year.
Other updates include a new developer feature called VizQL (Visual Query Language) Data Services, which allows enterprise users to embed Tableau anywhere in their automated business workflows.
“VizQL Data Service is a layer on top of published data sources and existing models, allowing developers to build composable data products using a simple programming interface,” the company said. increase.
Salesforce wins new users with Tableau-generated AI
In general, the addition of generative AI capabilities to Tableau can be seen as an attempt to attract customers who are not analytics or data experts. At least 35% of employees are unwilling to learn about analytics and data structures, Park said, so business intelligence suites face adoption issues.
“To get past that, analytics requires a radically different user interface. This combination of natural language processing, natural language generation, generative AI, and non-jargon input that transforms standard language into data relations provides that user interface,” added Park.
Another reason why new features are attractive to customers is business users’ indifference to using dashboards. “These users prefer to use natural language with context. Until now, NLP has been very difficult for computers to process, but his new LLM has changed that,” he said. added Mohan.
