Hi - Is it possible or planned to allow Sisense Narratives to connect to a different LLM provider? We are hosted in AWS and have restrictions around data transmission, so would need to use an internally available provider such as Bedrock instead of Arria for NLG. ... View more

Current behavior of the Simply Ask feature doesn't support automatic saving of the modified NLQ model to the "*.dash" file. So you loose your development work after exporting the dashboard with already enabled and tuned "Simply Ask" to a "*.dash" file. After importing this dashboard back to Sisense the feature is automatically disabled. Also if a developer will enable it manually it starts running based on the "By Default" NLQ model. This forces to redo all the work on the model back again which is time consuming and not efficient. We expect Sisense to consider in the future the development work on making this more convenient for a BI engineer. If a "Simply Ask" feature will be automatically enabled on an imported dashboard with a corrected model, it would be beneficial for dashboard deploy automation and will reduce unneeded additional manual work on setting up this feature again and again for different working environments for the same dashboard. ... View more

We create a basic set of pre-built dashboards that we publish, where business users can view but not modify.  We also have two-tiers of analytics offering for business users (standard and premium).  On our common, pre-built dashboards, which are available to both standard and premium, we would like the ability to expose Simply Ask to only those users of premium (hiding it from those who are standard).   While this may be achievable via a 2-model, 2-dashbaord set approach, we desire to manage a single set of dashbaords and model to simplify deployment overhead and maintenance. ... View more

It's the standard in our industry that products are on sale only for a limited time. For example, tickets being sold for an event aren't sold after the event has occurred. But there's not way to forecast future sales and prevent the forecast from predicting sales after the event has passed and no sales can possibly happen. Instead, customers have to edit every forecast that is currently predicting beyond its event date, every day, to predict only the correct number of future periods.  We'd like a way to predict N future periods, but not past a point in time at which inventory is scheduled to be no longer available. ... View more