Google generative - using response variable to update input_text helper

Configuration Frontend
1

I’m having a lot of fun with the Google AI stuff and my home cameras. I’ve successfully got an automation running that takes three snaps of a camera, and asks google to describe what has / is happening.

It then outputs the text as a response variable. I know this is working because I can then automate a notification to my iPhone that has the correct output of text.

What I’d also like to do is have this same output of text appear in a ‘input_text’ helper that I have setup so that I can see the last motion detected as a text box in my dashboard. This -for some reason- isn’t working. It will not update the text.

I’ve included the script below - can anyone see where I’ve gone wrong? Many thanks.

alias: Camera - Driveway 1 - Snapshot, AI & Notification (PERSON)
sequence:
  - alias: AWTRIX Person Outside
    data:
      topic: awtrix_efb7bc/notify
      payload: |-
        {
            "icon": "32198",
            "text": [
              {
                "t": "Person Outside",
                "c": "0000FF"
              }
            ],
            "pushIcon": 0,
            "repeat": 4
          }
    action: mqtt.publish
  - metadata: {}
    data:
      filename: ./www/snapshots/driveway1_personsnapshot1.jpg
    target:
      entity_id:
        - camera.g4_front_high
    enabled: true
    action: camera.snapshot
  - delay:
      hours: 0
      minutes: 0
      seconds: 1
      milliseconds: 0
    enabled: true
  - metadata: {}
    data:
      filename: ./www/snapshots/driveway1_personsnapshot1.jpg
    target:
      entity_id:
        - camera.g4_front_high
    enabled: true
    action: camera.snapshot
  - delay:
      hours: 0
      minutes: 0
      seconds: 2
      milliseconds: 0
    enabled: true
  - metadata: {}
    data:
      filename: ./www/snapshots/driveway1_personsnapshot2.jpg
    target:
      entity_id:
        - camera.g4_front_high
    enabled: true
    action: camera.snapshot
  - delay:
      hours: 0
      minutes: 0
      seconds: 2
      milliseconds: 0
    enabled: true
  - metadata: {}
    data:
      filename: ./www/snapshots/driveway1_personsnapshot3.jpg
    target:
      entity_id:
        - camera.g4_front_high
    enabled: true
    action: camera.snapshot
  - delay:
      hours: 0
      minutes: 0
      seconds: 1
      milliseconds: 0
    enabled: true
  - metadata: {}
    data:
      prompt: >-
        Motion has been detected caused by a person. Compare and very briefly
        describe what you see in the following images from my CCTV camera. You
        can arrange the screenshots in time order to make it easier. What do you
        think caused the motion alarm? Describe any people in detail. Do not
        describe stationary objects or buildings. Ensure output is 255
        characters or less.
      image_filename:
        - ./www/snapshots/driveway1_personsnapshot1.jpg
        - ./www/snapshots/driveway1_personsnapshot2.jpg
        - ./www/snapshots/driveway1_personsnapshot3.jpg
    response_variable: generated_content_p
    action: google_generative_ai_conversation.generate_content
  - action: input_text.set_value
    target:
      entity_id: input_text.google_ai_generated_text
    data:
      value: "{{generated_content_p['text'] }}"
  - metadata: {}
    data:
      title: Front - Person Detected
      message: "{{generated_content_p['text'] }}"
      data:
        image: /local/snapshots/driveway1_personsnapshot2.jpg
    action: notify.mobile_app_matts_iphone_15
  - action: tts.speak
    metadata: {}
    data:
      cache: true
      media_player_entity_id: media_player.kitchen_sonos
      message: "{{ generated_content_p.text }}"
    target:
      entity_id: tts.home_assistant_cloud
    enabled: false
mode: single
description: ""
icon: mdi:motion-sensor