Deployment Time: 4:30 AM PST
Features & enhancements
| Item
|
Description
|
| Feature
|
GA release of Extract rows and Summarize AI tasks in automations
Extract information from an uploaded image (such as a photo) or a file (PDF), and save it to rows in an existing table using AI by adding the Extract rows AI task in an automation. See Use the Extract rows AI task in an automation.
Summarize long textual data into a shorter summary using AI by adding the Summarize AI task in an automation. See Use the Summarize AI task in an automation.
|
Bug fixes
None
Rollout changes
None
Preview announcements
Preview feature releases enable you to try out new app features that are not yet fully supported. See Product launch stages.
- No new preview features were released today.
In addition, the AppSheet preview program lets app creators try out new app features that are not yet fully supported. For more details and to opt-in, see AppSheet preview program.
5 Likes
@lizlynch
Could you add the max allowable file size limit (presumably something like a 2.1 mb only…..) to your official documentation so that the app creators can be aware of this hard (low) limit?
Anyway, I was disappointed to see only max 12 columns being supported as output column.
Using Gemini/AI Studio API, we could derive and extract the much higher number of the column value out of the files/images, hence the limitation of only 12 columns is way too low to accommodate the real-world needs and use cases.
5 Likes
On top of my previous request, during our testing before, we realize the column with TIME type was not supported as output column. It is possible AI Task may not support other particular data type such as TIME/Datetime\Duration etc.
Could you add the list of supported data type as output (to the contrary, the list of not-supported data types) onto the official documentation so that the app creator will not lose a way during the configs. Thanks.
3 Likes
The file size limit will be updated to 12 MB this week. Thank you for the callout about adding it in docs. @lizlynch will update the docs with the new limit.
@Koichi_Tsuji How many columns are you trying to extract? We have to balance the number of columns with output reliability, consistency and cost.
Is there a particular reason you are unable to chain multiple extract tasks in an automation process (assuming you are using extract and not extract rows)?
3 Likes
It is very much difficult to say how many max number of the columns would be felt comfortable, but it is definitely too short to restrict with max 12 columns.
Any possible forms / template the app creators wish to extract should be complex ones, probably the number of the columns for those forms should be in the range of 50 - 100 columns (sorry the range is broad).
2 Likes
We have GAS, calling AI studio API (free), and extract the data out of PDF/Images, but it returns the reliable outputs without much of the Hallucination.
AppSheet is using the lower scale/capable version of Gemini to reduce the cost?
3 Likes
Let me share my feedback and general feeling to the newly GA-ed feature of “Extract Rows”.
The level of accuracy and reliability is far low and the actual/real world use cases could not be satisfied as it does not read a relatively (a bit) complex template. During our testing, we saw error message like this after running Extract Rows task number of the times.
Error Message
Error encountered in step with name [New step]: Error: AppSheet AI Service was unavailable: Object reference not set to an instance of an object.
On the face of app, the user could not see error message, but the error is only accessible through the monitoring app which is also not user-friendly. User not sure what happened to the jobs after posting files to the apps.
Interestingly, I posted those files Extract Rows task failed onto the AI Studio/Gemini Chat app and simply asked AI “please extract XXX, YYY, ZZZ values and display as table. This works like a charm, the AI returned mostly accurate values out of input files. Very strange, why AppSheet AI repeatedly fails while we believe the same AI Engine is behind the scene.
Further testing showed the gaps between Extract/Extract Rows tasks. For those files failed with Extract Row task, then switched to existing “Extract” task to read the values. This BOT (task) did not returned the error at all and presented the extracted values with good reliability and level of the accuracy. Not sure why Extract Rows task mostly fails while Extract task works fine.
Overall, the usage of Extract Rows task seems to be pretty much limited, only satisfy the very much straight ward and simple story as use cases.
@takuya_miyai
2 Likes
I rebuilt and tested the automation with Extract Rows task. And tested with same files (images) which have been throwing the error as I have shared and reported in the previous thread. It appears the works have been made by the Google teams and I could manage to extract the rows with success. This is a pretty much good news to us. Thanks.
3 Likes