Select to view content in your preferred language

Survey123 analysis tools (beta)

1887
2
07-01-2025 04:21 PM
MarikaVertzonis
Esri Regular Contributor
7 2 1,887

AI assistants in Survey123 are maturing. In a separate post you can read about the Survey123 assistant and how to use a conversation to help you design a survey and translate survey designs.

Also as exciting, is our progress in extracting information from images using natural language prompt (by generative AI). Over the last year our early adopters have provided great feedback about this capability, and it is now included in the Survey123 web designer as a beta feature for everyone to try.

Consider the following scenarios:

  • When recording a traffic violation, after capturing a photo of the vehicle, you want to enter the details of license plate number, vehicle color, and vehicle model into separate questions in the survey.
  • During a hydrant inspection, from the hydrant photo, you need to describe the state of the hydrant: in good condition, need maintenance, or is missing.

In both examples, the respondent could manually enter these details directly into the questions in the form themselves. But if many records are being captured, any opportunity to automate data entry is good, to minimize errors and to save time. The respondent can always correct the response that is generated, but human data validation is more efficient (and accurate) than human data entry.

Lets focus on the traffic violation scenario above. The survey author would configure each of the text questions in the survey to perform a calculation on the image question to extract the information required:

 

  1. In the web designer, choose to edit the calculation of the text question (say License plate number)

    MarikaVertzonis_0-1751410889250.png

  2. Select Vehicle image, and Analyze image from Extract property from the question

    MarikaVertzonis_1-1751410960382.png

  3.  Enter the prompt “Extract the license plate number from the vehicle”

    MarikaVertzonis_2-1751410988279.png

 

Once the survey is published, the web app will automatically populate the questions that have been configured using the assistant, once they capture a photo:

MarikaVertzonis_3-1751410994615.png

 

To use the image analysis tool (beta), the following must be configured by an ArcGIS organization administrator:

  • Disable the Block Esri apps and capabilities while they are in beta toggle. For more information see Blocked Esri apps.
  • Enable the Allow use of AI assistants by members of your organization toggle button. For more information see Allow AI assistants
  • Enable Analysis tools in the Survey123 website organization settings. For more information see Organization settings 

 

Not yet available on survey123.arcgis.com but under development, are audio transcription and text analysis from within the web designer. Before clicking the following link, be sure to register and sign into the Early Adopter Community.

Learn more about audio transcription and text analysis in Survey123.

Tags (3)
2 Comments
agx
by
Emerging Contributor

hello,

I created a survey123 with:
- an image element
- a text element calculated with Image Analysis: "What is this plant?"
I published the survey and shared it publicly. When I viewed it, I got the following message:This survey contains a feature that doesn't work anonymously.

thank you

Amanda__Huber
MVP Regular Contributor

@MarikaVertzonis is the "extract from the image" tool supported in the S123 App?