All of the players in the BI market are currently also focusing on adding AI related components to their suite. In this quarter’s blog post I’ll shortly go over what Tableau, Power BI and Qlik are currently working on.
- Tableau
Tableau is focusing on 2 features. Einstein copilot and Tableau Pulse.
Einstein copilot is basically a copilot that can help you with work you’re doing in Tableau. eg:
- “Recommended questions”: Generate a list of questions your connected data source could potentially answer.
- “Conversational data exploration”: You can give the copilot question you have on the data and it will try to generate visuals for you. Or, you can describe what you want to create and it will try to generate what you’ve described and in so saving you some time.
- “Guided calculation”: it can help you fix calculation syntax or even help you create calculations based on your description.
Tableau Pulse is powered by Einstein and can do the same things but is more focused towards the business users themselves and not the developers. Based on already defined metrics or metrics you create within the Pulse metric layer, users are able to subscribe to metrics they find important. Pulse will provide insights for these metrics in graph and text form and will try to give them questions (and answers to these questions) they might have regarding this metric. This will provide the users a different way to interact with the metrics they are interested in outside of the usual dashboard.
- Qlik Sense
For Qlik Sense it was a bit harder to find what they’re providing in terms of AI capabilities, since it’s not as marketed as other tools are marketing their AI capabilities. But Qlik Sense provides the same capabilities and more as what their competitors are providing. Through “Insight Advisor” Qlik is able to:
- Help users generate visuals based on natural language searches or selections of fields and existing master items.
- Chat based interface for conversational analytics that goes through all of the data and dashboards a user has access to and provides relevant answers and visualizations.
- Associative Insights to help uncovered blind spots or show relationships you may have missed.
- While creating visuals Qlik will give suggestions on what visual should be used.
To help the “Insight Advisor”, you’re able to (optionally) create a logical model of you data model and a vocabulary. Both of these will improve the success of the answers the tool will be able to provide to users.
Aside from the mentioned automated insight generation, search and natural language interaction, and AI-assisted creation and data prep. Qlik Sense also provides AutoML. This will basically allow non data scientists to also perform machine learning and predictive analyses in a no code way to analysts. Things like sales forecasting, churn reduction, customer acquisition, spend analysis, etc can now also be done from within Qlik.
- Power BI
Similar to Tableau, Power BI also has a copilot available to help developers in their work. Some of the features the copilot can help with:
- Help write DAX queries
- Help write your semantic model documentation
- Summarizes the content of your underlying semantic model, this can help you get a better understanding of you data.
- Based on questions and comments, suggest visuals or report outlines (what visuals, how many pages, what structure to use)
- Provide a summary of a visual you created or even a summary of your entire report. This provides you even more insights on what you can do with existing reports and visuals.
- …
As a report user (soon) you will also be able to ask copilot questions like:
- Summarize visuals or report pages.
- Ask questions about the data, copilot will try to answer you question in text format.
- Conclusion
As you see most of the current features will focus on mainly 3 things:
- A copilot which helps developers with creating reports and visuals
- Some kind of NLQ which helps users with questions they have. This can be from within a report/dashboard or even from outside of them (eg Tableau Pulse and Qlik).
- Predictive analysis by providing easy to use machine learning model to be applied on your data model.
Power BI is a bit more behind in the interaction possible by the users but I would think they will also eventually grow towards providing more functionality to the users. Making true ‘self service’ even more available as long as your data model is properly modeled and defined for the copilot to make the correct decisions.