Quick search…
Tutorials
Data integration
Support
Product Releases
2024-07-22
Instructions
The new instructions feature gives users the ability to configure the reponse style (e.g. tone, language); and provide additional context about the data or approach to specific analyses. This is broadly equivalent to having the ability to add additional commands into the system prompt, giving you a lot more control over the way the AI responds.
Chart table view & CSV download
Charts now include a "table view" as standard so you can see and download the underlying data (CSV) for processing in a spreadsheet or other data analysis tools.
2024-06-26
New nav bar
As well as some design improvements, we've done some spring cleaning to improved our navigation architecture. Now each of the "Data flow" sub-pages: Connections, Datasets and Metrics exists at the top level. This is to increase the visbility of the different configuration and optimization options, as well as prepare us for some upcoming feature developments.
Speed and accuracy improvements
We've doubled down on some speed improvements. Including switching to the new OpenAI ChatGPT 4o for parts of our pipeline.
2024-06-11
Flexible Insights
Flexible Insights, gives our AI the ability to coordinate multiple data queries and analyses to answer a single question.
This means we can answer more open-ended and complex questions. Some examples:
Prescriptive Questions
Which category of products should I focus my marketing efforts on?
What customer segments should be targeted in the next promotional campaign to ensure the highest ROI?
Diagnostic Questions
Why have my canceled orders increased in the last month? What are the patterns?
Which geographic locations yield the highest sales? What are the patterns?
Predictive Question
What will be the most popular product category next quarter based on past trends?
New homepage
We've simplified and beautified our homepage. We hope you like it!
Increased accuracy for Looker, Postgresql, Druid, Snowflake and Bigquery
We've made a combination of improvements to our data layer for the above connection types. Increasing question answering accuracy by around 20%!
2024-06-06
Better handling of categorical data types
We have improved our AI's understanding of categorical data types by providing it more context about the data itself. This means that questions like: "Which product is most popular in the USA" and "How many users completed the Gem Rush event last week" perform significantly better.
Role updates
There are now 4 standard roles:
Owner (complete access, billing, user management)
Project Admin (access to limited projects, user management)
Editor (can configure data within projects)
Questioner (can ask questions and view data)
Project level authorization
Users can now be granted access to only specific projects as needed.
Chart rendering improvements
We've made some optimizations to our chart rendering to improve speed and accuracy.
Description generation has been improved
We've made improvements to the description generation algorothm to include more context during the generation. This will help the AI have a better understanding of your data when you ask questions.
2024-04-30
Show Approach
Our new show approach enables you to see, validate and save the queries executed by the system. Click on “show approach” at the bottom of the answer to see, clearly, how the AI approached answering the question.
Clicking on a metric (seen underlined) will show you the specific query generated by the system. This definition can be saved, named and annotations for use going forwards. This helps you improve the AI for your organisation. Making the AI more consistent and strengthening its knowledge of your domain and business.
Improved charting and data retrieval
We are constantly improving and working on our accuracy and ability to answer questions. We’ve made a number of improvements to the effectiveness of the data retrieval and charting based on your feedback (i.e. negatively rated questions, and questions automatically marked as a miss) in the system.
Some specific highlights:
Better accuracy when applying a filter
Better accuracy with time ranges
Better use of aggregations
Improved accuracy when selecting metrics
Some edge cases where charts would fail to draw
Speed improvements
The whole system has gotten faster. We have reduced the average answer time by ~ 36%.
2024-07-22
Instructions
The new instructions feature gives users the ability to configure the reponse style (e.g. tone, language); and provide additional context about the data or approach to specific analyses. This is broadly equivalent to having the ability to add additional commands into the system prompt, giving you a lot more control over the way the AI responds.
Chart table view & CSV download
Charts now include a "table view" as standard so you can see and download the underlying data (CSV) for processing in a spreadsheet or other data analysis tools.
2024-06-26
New nav bar
As well as some design improvements, we've done some spring cleaning to improved our navigation architecture. Now each of the "Data flow" sub-pages: Connections, Datasets and Metrics exists at the top level. This is to increase the visbility of the different configuration and optimization options, as well as prepare us for some upcoming feature developments.
Speed and accuracy improvements
We've doubled down on some speed improvements. Including switching to the new OpenAI ChatGPT 4o for parts of our pipeline.
2024-06-11
Flexible Insights
Flexible Insights, gives our AI the ability to coordinate multiple data queries and analyses to answer a single question.
This means we can answer more open-ended and complex questions. Some examples:
Prescriptive Questions
Which category of products should I focus my marketing efforts on?
What customer segments should be targeted in the next promotional campaign to ensure the highest ROI?
Diagnostic Questions
Why have my canceled orders increased in the last month? What are the patterns?
Which geographic locations yield the highest sales? What are the patterns?
Predictive Question
What will be the most popular product category next quarter based on past trends?
New homepage
We've simplified and beautified our homepage. We hope you like it!
Increased accuracy for Looker, Postgresql, Druid, Snowflake and Bigquery
We've made a combination of improvements to our data layer for the above connection types. Increasing question answering accuracy by around 20%!
2024-06-06
Better handling of categorical data types
We have improved our AI's understanding of categorical data types by providing it more context about the data itself. This means that questions like: "Which product is most popular in the USA" and "How many users completed the Gem Rush event last week" perform significantly better.
Role updates
There are now 4 standard roles:
Owner (complete access, billing, user management)
Project Admin (access to limited projects, user management)
Editor (can configure data within projects)
Questioner (can ask questions and view data)
Project level authorization
Users can now be granted access to only specific projects as needed.
Chart rendering improvements
We've made some optimizations to our chart rendering to improve speed and accuracy.
Description generation has been improved
We've made improvements to the description generation algorothm to include more context during the generation. This will help the AI have a better understanding of your data when you ask questions.
2024-04-30
Show Approach
Our new show approach enables you to see, validate and save the queries executed by the system. Click on “show approach” at the bottom of the answer to see, clearly, how the AI approached answering the question.
Clicking on a metric (seen underlined) will show you the specific query generated by the system. This definition can be saved, named and annotations for use going forwards. This helps you improve the AI for your organisation. Making the AI more consistent and strengthening its knowledge of your domain and business.
Improved charting and data retrieval
We are constantly improving and working on our accuracy and ability to answer questions. We’ve made a number of improvements to the effectiveness of the data retrieval and charting based on your feedback (i.e. negatively rated questions, and questions automatically marked as a miss) in the system.
Some specific highlights:
Better accuracy when applying a filter
Better accuracy with time ranges
Better use of aggregations
Improved accuracy when selecting metrics
Some edge cases where charts would fail to draw
Speed improvements
The whole system has gotten faster. We have reduced the average answer time by ~ 36%.
2024-07-22
Instructions
The new instructions feature gives users the ability to configure the reponse style (e.g. tone, language); and provide additional context about the data or approach to specific analyses. This is broadly equivalent to having the ability to add additional commands into the system prompt, giving you a lot more control over the way the AI responds.
Chart table view & CSV download
Charts now include a "table view" as standard so you can see and download the underlying data (CSV) for processing in a spreadsheet or other data analysis tools.
2024-06-26
New nav bar
As well as some design improvements, we've done some spring cleaning to improved our navigation architecture. Now each of the "Data flow" sub-pages: Connections, Datasets and Metrics exists at the top level. This is to increase the visbility of the different configuration and optimization options, as well as prepare us for some upcoming feature developments.
Speed and accuracy improvements
We've doubled down on some speed improvements. Including switching to the new OpenAI ChatGPT 4o for parts of our pipeline.
2024-06-11
Flexible Insights
Flexible Insights, gives our AI the ability to coordinate multiple data queries and analyses to answer a single question.
This means we can answer more open-ended and complex questions. Some examples:
Prescriptive Questions
Which category of products should I focus my marketing efforts on?
What customer segments should be targeted in the next promotional campaign to ensure the highest ROI?
Diagnostic Questions
Why have my canceled orders increased in the last month? What are the patterns?
Which geographic locations yield the highest sales? What are the patterns?
Predictive Question
What will be the most popular product category next quarter based on past trends?
New homepage
We've simplified and beautified our homepage. We hope you like it!
Increased accuracy for Looker, Postgresql, Druid, Snowflake and Bigquery
We've made a combination of improvements to our data layer for the above connection types. Increasing question answering accuracy by around 20%!
2024-06-06
Better handling of categorical data types
We have improved our AI's understanding of categorical data types by providing it more context about the data itself. This means that questions like: "Which product is most popular in the USA" and "How many users completed the Gem Rush event last week" perform significantly better.
Role updates
There are now 4 standard roles:
Owner (complete access, billing, user management)
Project Admin (access to limited projects, user management)
Editor (can configure data within projects)
Questioner (can ask questions and view data)
Project level authorization
Users can now be granted access to only specific projects as needed.
Chart rendering improvements
We've made some optimizations to our chart rendering to improve speed and accuracy.
Description generation has been improved
We've made improvements to the description generation algorothm to include more context during the generation. This will help the AI have a better understanding of your data when you ask questions.
2024-04-30
Show Approach
Our new show approach enables you to see, validate and save the queries executed by the system. Click on “show approach” at the bottom of the answer to see, clearly, how the AI approached answering the question.
Clicking on a metric (seen underlined) will show you the specific query generated by the system. This definition can be saved, named and annotations for use going forwards. This helps you improve the AI for your organisation. Making the AI more consistent and strengthening its knowledge of your domain and business.
Improved charting and data retrieval
We are constantly improving and working on our accuracy and ability to answer questions. We’ve made a number of improvements to the effectiveness of the data retrieval and charting based on your feedback (i.e. negatively rated questions, and questions automatically marked as a miss) in the system.
Some specific highlights:
Better accuracy when applying a filter
Better accuracy with time ranges
Better use of aggregations
Improved accuracy when selecting metrics
Some edge cases where charts would fail to draw
Speed improvements
The whole system has gotten faster. We have reduced the average answer time by ~ 36%.
2024-07-22
Instructions
The new instructions feature gives users the ability to configure the reponse style (e.g. tone, language); and provide additional context about the data or approach to specific analyses. This is broadly equivalent to having the ability to add additional commands into the system prompt, giving you a lot more control over the way the AI responds.
Chart table view & CSV download
Charts now include a "table view" as standard so you can see and download the underlying data (CSV) for processing in a spreadsheet or other data analysis tools.
2024-06-26
New nav bar
As well as some design improvements, we've done some spring cleaning to improved our navigation architecture. Now each of the "Data flow" sub-pages: Connections, Datasets and Metrics exists at the top level. This is to increase the visbility of the different configuration and optimization options, as well as prepare us for some upcoming feature developments.
Speed and accuracy improvements
We've doubled down on some speed improvements. Including switching to the new OpenAI ChatGPT 4o for parts of our pipeline.
2024-06-11
Flexible Insights
Flexible Insights, gives our AI the ability to coordinate multiple data queries and analyses to answer a single question.
This means we can answer more open-ended and complex questions. Some examples:
Prescriptive Questions
Which category of products should I focus my marketing efforts on?
What customer segments should be targeted in the next promotional campaign to ensure the highest ROI?
Diagnostic Questions
Why have my canceled orders increased in the last month? What are the patterns?
Which geographic locations yield the highest sales? What are the patterns?
Predictive Question
What will be the most popular product category next quarter based on past trends?
New homepage
We've simplified and beautified our homepage. We hope you like it!
Increased accuracy for Looker, Postgresql, Druid, Snowflake and Bigquery
We've made a combination of improvements to our data layer for the above connection types. Increasing question answering accuracy by around 20%!
2024-06-06
Better handling of categorical data types
We have improved our AI's understanding of categorical data types by providing it more context about the data itself. This means that questions like: "Which product is most popular in the USA" and "How many users completed the Gem Rush event last week" perform significantly better.
Role updates
There are now 4 standard roles:
Owner (complete access, billing, user management)
Project Admin (access to limited projects, user management)
Editor (can configure data within projects)
Questioner (can ask questions and view data)
Project level authorization
Users can now be granted access to only specific projects as needed.
Chart rendering improvements
We've made some optimizations to our chart rendering to improve speed and accuracy.
Description generation has been improved
We've made improvements to the description generation algorothm to include more context during the generation. This will help the AI have a better understanding of your data when you ask questions.
2024-04-30
Show Approach
Our new show approach enables you to see, validate and save the queries executed by the system. Click on “show approach” at the bottom of the answer to see, clearly, how the AI approached answering the question.
Clicking on a metric (seen underlined) will show you the specific query generated by the system. This definition can be saved, named and annotations for use going forwards. This helps you improve the AI for your organisation. Making the AI more consistent and strengthening its knowledge of your domain and business.
Improved charting and data retrieval
We are constantly improving and working on our accuracy and ability to answer questions. We’ve made a number of improvements to the effectiveness of the data retrieval and charting based on your feedback (i.e. negatively rated questions, and questions automatically marked as a miss) in the system.
Some specific highlights:
Better accuracy when applying a filter
Better accuracy with time ranges
Better use of aggregations
Improved accuracy when selecting metrics
Some edge cases where charts would fail to draw
Speed improvements
The whole system has gotten faster. We have reduced the average answer time by ~ 36%.
2024-07-22
Instructions
The new instructions feature gives users the ability to configure the reponse style (e.g. tone, language); and provide additional context about the data or approach to specific analyses. This is broadly equivalent to having the ability to add additional commands into the system prompt, giving you a lot more control over the way the AI responds.
Chart table view & CSV download
Charts now include a "table view" as standard so you can see and download the underlying data (CSV) for processing in a spreadsheet or other data analysis tools.
2024-06-26
New nav bar
As well as some design improvements, we've done some spring cleaning to improved our navigation architecture. Now each of the "Data flow" sub-pages: Connections, Datasets and Metrics exists at the top level. This is to increase the visbility of the different configuration and optimization options, as well as prepare us for some upcoming feature developments.
Speed and accuracy improvements
We've doubled down on some speed improvements. Including switching to the new OpenAI ChatGPT 4o for parts of our pipeline.
2024-06-11
Flexible Insights
Flexible Insights, gives our AI the ability to coordinate multiple data queries and analyses to answer a single question.
This means we can answer more open-ended and complex questions. Some examples:
Prescriptive Questions
Which category of products should I focus my marketing efforts on?
What customer segments should be targeted in the next promotional campaign to ensure the highest ROI?
Diagnostic Questions
Why have my canceled orders increased in the last month? What are the patterns?
Which geographic locations yield the highest sales? What are the patterns?
Predictive Question
What will be the most popular product category next quarter based on past trends?
New homepage
We've simplified and beautified our homepage. We hope you like it!
Increased accuracy for Looker, Postgresql, Druid, Snowflake and Bigquery
We've made a combination of improvements to our data layer for the above connection types. Increasing question answering accuracy by around 20%!
2024-06-06
Better handling of categorical data types
We have improved our AI's understanding of categorical data types by providing it more context about the data itself. This means that questions like: "Which product is most popular in the USA" and "How many users completed the Gem Rush event last week" perform significantly better.
Role updates
There are now 4 standard roles:
Owner (complete access, billing, user management)
Project Admin (access to limited projects, user management)
Editor (can configure data within projects)
Questioner (can ask questions and view data)
Project level authorization
Users can now be granted access to only specific projects as needed.
Chart rendering improvements
We've made some optimizations to our chart rendering to improve speed and accuracy.
Description generation has been improved
We've made improvements to the description generation algorothm to include more context during the generation. This will help the AI have a better understanding of your data when you ask questions.
2024-04-30
Show Approach
Our new show approach enables you to see, validate and save the queries executed by the system. Click on “show approach” at the bottom of the answer to see, clearly, how the AI approached answering the question.
Clicking on a metric (seen underlined) will show you the specific query generated by the system. This definition can be saved, named and annotations for use going forwards. This helps you improve the AI for your organisation. Making the AI more consistent and strengthening its knowledge of your domain and business.
Improved charting and data retrieval
We are constantly improving and working on our accuracy and ability to answer questions. We’ve made a number of improvements to the effectiveness of the data retrieval and charting based on your feedback (i.e. negatively rated questions, and questions automatically marked as a miss) in the system.
Some specific highlights:
Better accuracy when applying a filter
Better accuracy with time ranges
Better use of aggregations
Improved accuracy when selecting metrics
Some edge cases where charts would fail to draw
Speed improvements
The whole system has gotten faster. We have reduced the average answer time by ~ 36%.
Support
Product Releases
2024-07-22
Instructions
The new instructions feature gives users the ability to configure the reponse style (e.g. tone, language); and provide additional context about the data or approach to specific analyses. This is broadly equivalent to having the ability to add additional commands into the system prompt, giving you a lot more control over the way the AI responds.
Chart table view & CSV download
Charts now include a "table view" as standard so you can see and download the underlying data (CSV) for processing in a spreadsheet or other data analysis tools.
2024-06-26
New nav bar
As well as some design improvements, we've done some spring cleaning to improved our navigation architecture. Now each of the "Data flow" sub-pages: Connections, Datasets and Metrics exists at the top level. This is to increase the visbility of the different configuration and optimization options, as well as prepare us for some upcoming feature developments.
Speed and accuracy improvements
We've doubled down on some speed improvements. Including switching to the new OpenAI ChatGPT 4o for parts of our pipeline.
2024-06-11
Flexible Insights
Flexible Insights, gives our AI the ability to coordinate multiple data queries and analyses to answer a single question.
This means we can answer more open-ended and complex questions. Some examples:
Prescriptive Questions
Which category of products should I focus my marketing efforts on?
What customer segments should be targeted in the next promotional campaign to ensure the highest ROI?
Diagnostic Questions
Why have my canceled orders increased in the last month? What are the patterns?
Which geographic locations yield the highest sales? What are the patterns?
Predictive Question
What will be the most popular product category next quarter based on past trends?
New homepage
We've simplified and beautified our homepage. We hope you like it!
Increased accuracy for Looker, Postgresql, Druid, Snowflake and Bigquery
We've made a combination of improvements to our data layer for the above connection types. Increasing question answering accuracy by around 20%!
2024-06-06
Better handling of categorical data types
We have improved our AI's understanding of categorical data types by providing it more context about the data itself. This means that questions like: "Which product is most popular in the USA" and "How many users completed the Gem Rush event last week" perform significantly better.
Role updates
There are now 4 standard roles:
Owner (complete access, billing, user management)
Project Admin (access to limited projects, user management)
Editor (can configure data within projects)
Questioner (can ask questions and view data)
Project level authorization
Users can now be granted access to only specific projects as needed.
Chart rendering improvements
We've made some optimizations to our chart rendering to improve speed and accuracy.
Description generation has been improved
We've made improvements to the description generation algorothm to include more context during the generation. This will help the AI have a better understanding of your data when you ask questions.
2024-04-30
Show Approach
Our new show approach enables you to see, validate and save the queries executed by the system. Click on “show approach” at the bottom of the answer to see, clearly, how the AI approached answering the question.
Clicking on a metric (seen underlined) will show you the specific query generated by the system. This definition can be saved, named and annotations for use going forwards. This helps you improve the AI for your organisation. Making the AI more consistent and strengthening its knowledge of your domain and business.
Improved charting and data retrieval
We are constantly improving and working on our accuracy and ability to answer questions. We’ve made a number of improvements to the effectiveness of the data retrieval and charting based on your feedback (i.e. negatively rated questions, and questions automatically marked as a miss) in the system.
Some specific highlights:
Better accuracy when applying a filter
Better accuracy with time ranges
Better use of aggregations
Improved accuracy when selecting metrics
Some edge cases where charts would fail to draw
Speed improvements
The whole system has gotten faster. We have reduced the average answer time by ~ 36%.