Questions and answers are generated from Webex chat and transcrip
Q: What is the difference between the “Create” button versus the “Generate Analysis” button?
A: The “Create” button is the conventional create analysis button — when you click it, you land on an empty canvas and then select fields, visualizations, and do everything manually. The “Generate Analysis” button is an add-on feature that lets you describe what you want in natural language and generates a pre-built analysis for you. All authors are getting promotional access to Generate Analysis until the end of this year across all seven Amazon Quick regions, but over time it will only be available for Author Pro licenses.
Q: Can this feature be used with data sets in my own account instead of the Amazon BI account shown in the demo?
A: Yes, this will work on your account. The Amazon BI account shown in the demo is just an internal account used for demonstrations. Generate Analysis works on all your accounts, all your existing data sets — everything works as-is. It is simply a new way of creating your dashboard.
Q: My team’s biggest blocker is access to various data sets relevant to our work. Are there any tips on securing permissions or access in bulk or individually?
A: Check out Amazon Quick Spaces. Spaces is a feature that lets you create team-specific spaces and organize all the information, data sets, dashboards, and everything into those spaces. Spaces is a great way to organize data in bulk for your entire team so you can give them access all at once.
Q: As part of the product roadmap, do we have plans where we don’t necessarily have to create a topic on the dataset and have a chat agent in Quick that can directly link to the data set where the user can start asking Q&A?
A: Yes! There is a feature coming very soon where you can directly ask questions on your data set without needing to create anything — you don’t even need to create a dashboard. If you have a data set, you can start asking questions and it will answer them. Stay tuned for the May 5th Amazon Quick Learning Series session where our solution architects will be covering this and other new features.
Q: How does Amazon Quick Desktop’s local file indexing work with my existing spaces and data sets?
A: The indexing works with native technologies including OpenSearch, along with different indexing methods used to index your data alongside your data and spaces to provide comprehensive answers. For a more detailed answer, we recommend reading the documentation online, and we will also be covering Amazon Quick Desktop in a dedicated session later in May.
Q: Does the Generate Analysis feature work if I want to create geospatial dashboards? My personal experience is that mapping to geospatial fields only works in the old experience.
A: Currently, Generate Analysis does not support geospatial visualizations. However, geospatial charts, Highcharts, and similar chart types are coming very soon as part of the feature roadmap.
Q: Once a dashboard is generated from these data sets, can I access and interact with it through other forms of Quick like Quick Desktop? Does the agent in Desktop have the same ability to answer questions on the dashboard as the web browser experience?
A: Yes. Once the analysis is created and published as a dashboard, it becomes part of your overall assets. You go through the same publishing flow, and then you can start asking questions on it. Quick Desktop will also recognize the dashboard — if you name it and start asking questions from that particular dashboard, it will work. If the dashboard is part of any space, and that space is something you’re querying, it will consider this dashboard as well.
Q: Can I ask to have tooltips added during the Generate Analysis and will it successfully build that?
A: That capability is coming soon. Advanced tooltips, advanced interactions, highlights, adding a new sheet using a prompt, and editing everything through a common chat functionality are all part of the roadmap and will be available soon.
Q: How do I improve the dashboard using Generate Analysis? Can I use it multiple times to improve it? If so, are each iteration versioned?
A: Currently, Generate Analysis creates the analysis one time, and then you publish it as a dashboard. If you want another version, you go back and refine your prompt based on what you learned. However, the ability to regenerate a sheet or regenerate the entire analysis is coming soon. Note that analyses are not versioned entities — dashboards are versioned (each publish creates a new version you can roll back to). Analyses have an undo/redo stack, so you can undo any manual changes you make after generation.
Q: What is “Enable Quick Actions”? I saw a checkbox that showed up when the analysis is published.
A: The “Enable Quick Actions” checkbox controls whether you want actions in automations or Quick Flows to be taken based on the data in this dashboard. If you don’t want any automated agent to take actions based on the dashboard data, keep it unchecked. If you’re okay with automated actions — for example, getting alerted via Slack or Outlook email when a threshold is reached or a metric goes above a certain limit — you can check this box to enable it.
Q: Will Generate Analysis work with direct query data sets?
A: Yes, it will work with direct query data sets, but SPICE is recommended because the speed of generation will be significantly faster. With direct query, the speed will be much slower.
Q: Can Quick Desktop generate analyses or visuals from data sets, or is that still just for the web app?
A: Right now, Generate Analysis is only available on the web app. We will be covering Amazon Quick Desktop in a dedicated session later in May.
Q: Can I use Quick Chat to help create the prompt for generating the dashboard if it’s a complex multi-step request?
A: Yes, you can! Some early beta authors have already been doing this. You can use Quick Chat to talk to your data, ask for ideas on what kind of visualizations to create, and even ask it to help write a prompt. You can do all kinds of prompt engineering in the chat, then copy and paste that prompt into Generate Analysis.