A "data" sign at Tableau's conference in San Diego, California
Tableau unveiled new data tools for the rest of us in San Diego yesterday, mostly powered by AI. Hyped-up product developers showed ways to plan vacations with data almost automatically or create complex-looking visualizations of data about the art they love with just a few clicks. Executives talked about APIs and shared dimensions and composable data sources and new integrations with Microsoft Teams, plus how “data helps us make the world a better place.”
But the real innovation data visualization giant Tableau (owned by Salesforce) is pushing is instant access to data for everyone, just by asking, via advanced AI technologies similar to those that power ChatGPT.
“Today's tools are like hammers: you have to do it yourself,” Tableau chief product officer Southard Jones told me yesterday. “They’re not a nail gun that does it for you.”
Building that nail gun is no easy task in today’s data-rich organizations. One that I talked to which is using one of the new Tableau AI innovations, called Einstein Copilot, says that his organization built a data lake with over 30 terabytes of data: too much to be accessible by any standard, quick, or inexpensive means. After creating snapshots of the data in Snowflake, the company enabled access via Copilot.
“We can mine the data now, we can see trends,” says Brim Basom, a managing director at Goosehead Insurance. “Now I can see truly what’s happening in the organization, and we’re enabling our end users to have better tools.”
In addition to Einstein Copilot, aimed largely at data analysts and data scientists, Tableau revealed more details about Pulse, the product for the rest of us that was originally teased back in early 2023. The goal with Pulse is to simply ask your data questions in natural language: how much of our budget have we spent? what are the results for our latest marketing campaign? why are we out of stock of Widget 357?
But there is a significant challenge that precedes data insights. Better access to data is not the full solution, if the base data itself is flawed.
“Problems start upstream,” says Walid Mehanna, the chief data & AI officer for international science and technology giant Merck KGaA. “We treat data as a waste product.”
One of the fixes for that which he’s working on? You guessed it: more AI.
“Why do I have people put in data manually?” Mehanna asks. “I want a system to generate good quality data.”
In other words, multiple data-creating parts of the enterprise such as sales and customer service create vast quantities of highly valuable insight, if it can be mined. But often it’s unstructured, or if it is structured, it is categorized in the wrong way thanks to old systems with inefficient data entry mechanisms. People are trying to do the right thing, but they also need to get their jobs done, and often it’s easier to kind of hack the system to get your job done rather than ensure every piece of data entered is done by the letter of the law.
Mehanna says Merck is working on its own AI technology to allow service, sales, and other internal staff to enter data simply by asking a smart system to do it. He’s launched MyGPT at Merck, using a German startup’s technology as the platform, to build that component.
But he does see huge value in getting quality data insight out, once the base data is accurate.
“The vision is powerful,” he says. “Pulse is a paradigm shift, because reports and dashboards are too prevalent. What we need are actionable, timely insights: a system that suggests we look at this KPI because something is unusual.”
In other words, he’s looking for AI solutions to guide people’s attention so they don’t have to go through multiple reports or dashboards. And to highlight urgent but not immediately visible situations that otherwise might remained buried in the data.
For Tableau and Salesforce, the solution is clear, if not simple.
“Every company needs to focus on becoming an AI enterprise,” says Tableau CEO Ryan Aytay. “Get your data in order, find your single source of truth.”
That means getting the data right in the first place, then making it available to the people who need it easily and quickly, and finally translating that into action, whether that’s in another Salesforce acquisition, Slack, or Microsoft Teams, which Goosehead Insurance finds useful.
“Ultimately, it’s about leveraging AI in everything we do,” Aytay says.
One caution from Jones, Tableau’s product chief: it’s not magic. It’s technology, and that’s why we still need the people, the data experts alongside the business experts who understand what the data represent.
