Tableau has come a long way since it was acquired in mid-2019. Most of Tableau’s advancements are already apparent to those who use or keep up to date with Tableau. However, with Tableau Einstein announced at Dreamforce, it’s not a case of another product renaming – this announcement has some serious significance.
We had the chance to sit down with Tableau CEO, Ryan Aytay, who was cheerful despite it being the last interview from his long run of meetings at Dreamforce. Our conversation covered why Tableau Einstein is a huge milestone for Tableau, when we can expect it to arrive “into our hands”, predictions for Tableau’s future, and much more.
Understanding Tableau Einstein
Q: What is Tableau Einstein and how does it impact Tableau’s capabilities and presence within the Salesforce ecosystem?
“Tableau Einstein is a completely new platform that’s built on the Salesforce ‘core’ platform. It’s been built with Data Cloud as the data foundation, and has Agentforce embedded.
“It’s exciting because it allows Tableau to go beyond being focused only on visualization, while still keeping hold of Tableau’s original mission: to help people see and understand data.”
Since Salesforce acquired Tableau around five years ago, the vision has been to bring Tableau technology and Salesforce technology closer together so that we can take the best of both worlds.
The Salesforce platform gives a robust foundation, and with Tableau more integrated, it allows Tableau to start appearing in different parts of the Customer 360 (Sales Cloud, Service Cloud, etc.)
Why Data Cloud?
The world has changed, and AI is producing and using more data daily. While people still want to visualize data, they also have data issues across their business – this is why Data Cloud or other data layers are required.
Tableau Copilot → Tableau Einstein + Agentforce
Q: Generative AI is improving fast, and as a result, tasks and data management will become more automated. So, what’s the big splash that differentiates Tableau Copilot from the Tableau Einstein and Agentforce duo?
“With the Copilots across the Salesforce platform, the user was guiding the AI – now, with agents, the user is still guiding the AI, but the AI is guiding them in return. As we look at Copilot in its first iteration, it was very much for the novice analyst.
“If you instruct the copilot to create a dashboard, it returns a data set and creates the visualization. Great – it’s helpful, but it wouldn’t be automatic.
“The next generation (or upgrade) focuses on the agent. While not absolutely autonomous, it reasons better and is far more assistive.”
Tableau Einstein vs. Tableau Suite
Q: What does the product suite now look like up until this point, and where is it heading? With the introduction of Tableau Einstein, could more products be added to the suite?
“The product suite has become more broad over time, with the intention that it’s increasingly relevant for any data professionals.
We already mentioned that Tableau has been rebuilt and re-architected on the Salesforce core platform, with Data Cloud a part of it – resulting in Tableau Einstein.”
Tableau has gone from a files-based visualization tool* (what it was built for which led to 21 years of success), to now inheriting Salesforce’s modern UI, utilizing metadata from the Salesforce platform, and is faster and more secure.
*This is explained at the end of this article, in the section “Files-Based Visualization Tool Explained”.
If you have a lot of data in Tableau, you have to do an extract (i.e. grab data and pull it out of the data warehouse, do mining work, and then create a visualization), which is not a great process.
In the future, you can do live queries on Tableau at scale. “That is hard to explain and to showcase”, Ryan Aytay admitted, then went on to say “When you think about large companies, such as large banks and most of the federal government, their ability to scale while benefiting from better performance, and having usability, is phenomenal when you move to the new platform.”
The Path Toward Tableau Einstein: Consolidation and Enhancements
Salesforce doesn’t deny that they have multiple analytics offerings. This is a result of acquisitions, and how Salesforce have focused their product development. Consolidation is on the cards for the collection of products and features (listed below), with analytics coming into the Tableau Einstein environment.
Let’s take a look at these products and features:
CRM Analytics (CRMA)
The analytics product that has “grown up” organically when Salesforce acquired the technology (formerly Einstein Analytics, and formerly Wave). However, this is now considered a legacy product with the advent of Tableau Einstein.
“One of the downsides with CRMA is that you can’t click into the visualizations and deep dive like you can with Tableau,” Ryan says. “Our plan has been to take the front end of the user experience and pair it with Tableau’s exploration capabilities.”
“The combination means that we can feed these rich analytics into Sales Cloud, Service Cloud (etc.) UI. Over time, this will become the new CRMA.”
Tableau’s Visualization Suite
The Tableau Visualization Suite assigns users to one of four roles (granted permissions) according to how they work with the tooling (these are explorers. viewers and creators). The four roles are:
- Tableau Prep: The data management portion of Tableau. It provides a visual interface as you join datasets to make them usable for creating visualizations.
- Semantic layer: A translation layer, or a way to have a common business language around different data aspects.
- Workspaces: A sandbox to create and play with different datasets.
- Marketplace: Upload what’s been created in order to be shared.
Tableau Cloud
Tableau hosted over the web (the opposite of an on-premise or hybrid solution) to power their analytics. The need for speed is a driving factor.
Tableau Server
Tableau is hosted on-premise.
Tableau Pulse
Condenses down the UI to the most key metrics, relevant to you. AI anticipates the data each user needs based on personal preferences and automatically generates the insights.
Tableau Public
Provides functionality, freely available, therefore eliminating the cost implications and resource accessibility that may prohibit analysts from learning new skills.
Revenue Intelligence
Dedicated dashboards for sales teams, powered by CRMA that offer actionable insights and visibility. Revenue Insights and Sales Rep Command Center are examples within the product’s collection.
Service Intelligence
Dedicated dashboards for service teams, powered by CRMA that offer actionable insights and visibility into KPIs such as CSAT, average time to close, and total escalated cases.
To reiterate, this all will fall under Tableau Einstein. However, this is more than just a renaming – the Tableau Einstein name symbolizes the movement from multiple data/tech stacks and environments, into one.
A burning question that (hopefully) will no longer need to be asked is, “Should I buy CRMA or Tableau?”, as it will be one common, unified analytics solution (i.e. Tableau Einstein).
What About Tableau Customers That Don’t Use Salesforce?
Q: For Tableau customers that don’t have Salesforce or users that don’t have a Salesforce user license, can they still use the functionality that Tableau Einstein promises?
“Just because we’ve built, and will continue to build on the Salesforce platform, we are not a tool that is exclusively for CRM – in fact, that’s still a relatively small percentage of the user base.
“We frequently see customers that operate with a mix. There’s one Tableau customer who has a significant investment in Salesforce technology, including heavy usage of Data Cloud and Tableau. However, only around 20% of their Tableau touches Salesforce, with the rest of the users spread across other business departments, for example, in finance or HR.
“For those who wouldn’t have a Salesforce user license, the Tableau that they access would still be on feature parity. This is why we must be a platform that serves all needs.
“Then, there’s Tableau’s on-premise technology, a differentiator, in short, because it allows organizations to deploy the way they want to deploy. In this case, will every generative AI feature work? No, because AI-enabled features run in an LLM (external to the on-premise set-up). We have different ways to serve each customer, so having a flexible platform is a big part of Tableau’s ethos.”
Tableau Einstein Roll-Out
This will take place over the next few months, with estimated time frames “knock on wood because we don’t always know exactly” (Ryan Aytay).
- Phase 1: The first part of Tableau Einstein goes GA in October 2024. This will include Nonprofit Cloud and Marketing Cloud (the Growth and Advanced editions, built on Salesforce core).
- Phase 2: “When you’ll start to see Tableau Einstein light up” – will be February 2025. Around that time frame, we should see it applicable to Sales Cloud, Service Cloud, and Revenue Cloud.
Tableau’s Three-Step Mission with Tableau Einstein
1. Reusability
An issue in the data analysis industry is that you can’t really reuse what’s been built, and apply it to other use cases. For example, you create a travel and expense dashboard with a great data preparation flow and model around it – what if you want to share that with other departments in your organization? Traditionally, this hasn’t been easy.
The way the new platform is architected allows for templates – it’s like you can draw a box around certain components, and put it into a marketplace to be shared internally for other colleagues to take advantage of. It sounds somewhat simple, but it’s actually really powerful in our industry because it has never existed until now.
2. Collaboration
Day to day, Ryan spends time talking to Tableau customers, partners, and the community about what the future of Tableau should look like. Tableau Einstein still focuses on the core audience of Tableau – the Datafam, a group of passionate data people.
“When asking ourselves: How do we move to the future? Listening is a big part of it,” Says Ryan. “Much of my working days are consumed in bringing information to the community (“Here’s what we’re thinking, here’s what we’re building”), listening to the feedback, then adjusting how we’re planning to build and take the product to market.
“Aside from growing the business (i.e. selling more software), we are also showing existing and prospective customers the future”.
3. The “Agent” Concept
A lot of the future now is very focused on, obviously, AI. Every customer wants to talk about AI, but it now becomes more exciting due to the concept of agents.
An example is a travel and expense dashboard (from point #1). Almost all organizations care about cost, so, to enhance this visualization, an agent could be created that behind the scenes, should someone go over their daily threshold, the agent flags this and takes an action (e.g. notify via Slack). Agents automate the process of making decisions 24 hours a day.
There’s also the concept of the agent following your work. For example, you’ve created a geographic representation data and then moved into the Tableau Workspace. The agent will still show up in whichever medium of the Tableau suite.
Data Cloud Consumption for Tableau
Q: Some Salesforce products have demonstrated a movement from user seats to consumption-based pricing – as seen most apparently with Data Cloud (credits) and also Agentforce (baseline price $2 per conversation). So, with Tableau Einstein with these both, what will the cost of consumption look like?
“Tableau would take the same Data Cloud credit consumed. Say your organization already has Data Cloud, those Data Cloud credits can be applied to what you’re performing in Tableau Einstein. There’s no consumption per se when you run a visualization (or similar).”
Where credit consumption kicks in, your organization will increase the amount of data (“tons and tons”) processing through Data Cloud – that’s where the cost will incur. As Salesforce keeps saying, the aim is to simplify consumption-based pricing for customers.
Tableau Einstein Alliance
Salesforce just announced the Tableau Einstein Alliance, featuring established names like Deloitte and IBM, to help customers leverage AI through Tableau Einstein, integrated with Salesforce.
Partners will have early access to the platform, and insights on AI advancements and co-selling opportunities. Solutions from the Alliance will soon be available on the AppExchange and Tableau Marketplace, which should give developers the chance to innovate and monetize new AI-driven tools.
Predictions for Tableau’s Future
Q: It’s clear that Tableau Einstein is poised to redefine how companies approach analytics. Given the platform’s rapid evolution and its integration with Salesforce technologies, how do you envision Tableau Einstein shaping the future of BI tools over the next five years? And more importantly, what does this mean for companies looking to leverage AI-powered insights to drive growth and success?
“I like to compare it to Salesforce. I started working at Salesforce in 2006 when we were a one or two-product company – we had Sales Cloud and the platform (we were just getting into service).
“Likewise, Tableau was a one-product company, and its suite is now growing beyond, more broadly – the semantic layer, data warehousing type of experience with Data Cloud, and the marketplace are all examples. So, Tableau is becoming a multi-product suite that’s deeply integrated into Salesforce. Tableau ‘on core’ (Salesforce platform) will be a huge part of the five-year roadmap.
“There has always been enough demand to focus on our core audience (data professionals), and now we’re able to focus on a broader audience that also needs analytics – e.g. sales who need analytics but with no desire to click around too much to find the insight they want. By building Tableau Einstein into Salesforce, you ‘just turn it on’ – it’s a highly integrated, auto-on environment.
“There’s still Tableau the brand – it’s very powerful, and recognized in some cases differently from Salesforce. I predict a future where there are agents above agents, i.e. supervisory agents. We, at Salesforce, haven’t talked about this extensively yet, but it’s coming.”
Files-Based Visualization Tool Explained
When describing what Tableau has been (its 21 years of success), to what it is now, we mentioned that it has gone from a files-based* visualization tool to now being “re-architected” into the blossoming Tableau Einstein.
An example that’s easiest to understand uses Tableau Pulse (the Tableau offering that condenses down the UI to the most key metrics, relevant to you). When opening the app, you will see what’s called a metric, such as an open pipeline this quarter; and what’s called an insight which uses machine learning to enhance the metric.
When clicking into a metric, the data source from where the metric has been derived is displayed. This source is in fact a data file (a workbook that sits in Tableau). If you needed to query this data in the moment (i.e. ask a question about the data) and have the answer show up, the answer would be restricted to what can be gleaned from structured data.
Alternatively, think of the way it’s been architected as “blocks”, or using the familiar concept of spreadsheets, that organize data into rows and columns, contained in different workbooks.
Let’s consider the opposite of the picture painted above. Breaking away from queries based strictly on structured data, what if queries, and their subsequent answers could also consider unstructured data? This is the vision that’s possible with Data Cloud involved. With its Vector Database (that can process unstructured data) and Atlas’ RAG engine, more challenging queries can be asked of Tableau Einstein, and the answers retrieved faster (in real-time).
Summary
As Tableau Einstein continues to grow, we can expect even more amazing features and tools that will help it potentially transform into a multi-product suite that can tackle everything – from data prep to predictive modeling.
The post Conversation With Tableau CEO: Moving on to Core, Agents, and Future of CRM Analytics appeared first on Salesforce Ben.