Physical Intelligence

Physical Intelligence is part of Outset's Visual Intelligence Suite. It enables participants to upload photos during an interview — of their home, workspace, products, or surroundings — and lets the AI moderator analyze those images in real time to ask smarter, more contextual follow-up questions.

This is particularly useful when the research question lives in the physical world: what's in someone's pantry, what products are on their shelves, or how they interact with a device in their home.

Part 1: Guide Programming

Adding a Participant Upload Section

Physical Intelligence introduces a new section type in your guide called Participant Upload. To add one:

1. Open your guide and click + Add Section

1. Select Participant Upload from the section type menu

2. Write a prompt that tells participants what to upload (e.g., "Please take or upload a photo of the inside of your fridge")

Optionally, add…

1. text instructions to give participants more context or guidance. If the AI determines the uploaded image doesn't match your prompt, the participant will be asked to re-upload once before the interview continues regardless

💡 Tip: Be specific in your upload prompt. The more clearly you describe what you're looking for, the better the AI can validate the image and generate relevant follow-up questions.

Adding Questions to a Participant Upload Section

All standard question types are available within a Participant Upload section. Any questions placed after the upload step will automatically display the participant's uploaded image as a stimulus — just like a researcher-uploaded image — so participants can refer back to it while answering.

The AI moderator will have full awareness of the image content via a text summary generated by the silent observer, allowing it to ask intelligent, image-aware follow-ups.

Part 2: The Participant Experience

Uploading a Photo

When a participant reaches a Participant Upload section, they'll be presented with two options:

• Take a photo — use their device camera to capture a new one

• Upload a photo — choose an existing image from their device

Only one photo can be submitted per section.

Image Validation & Re-upload

Once a photo is submitted, the AI runs a quick check to determine whether the image matches the researcher's prompt. If it doesn't match, the participant will be prompted to try again with a clearer image. After one retry, the interview moves forward regardless of the result.

Note: During this phase, mismatched images are tracked internally and flagged on the Transcript and Insights page, but are not routed to Outset's fraud pipeline. This may change in a future release.

Part 3: Analysis

What you'll see in transcripts

Each uploaded image is treated as a piece of stimulus within the transcript — similar to a researcher-uploaded image. Alongside the image, you'll find:

• A text summary of the image generated by the AI, giving you a quick read on what was captured

• Item tags — automatically generated labels identifying objects and elements visible in the image (e.g., "cereal box," "cleaning products," "dog")

Aggregate view

Item tags are surfaced both for individual responses and in aggregate across all participants — making it easy to spot patterns, like which products appear most frequently across your sample.

Example use cases

FAQs

Physical Intelligence works with text, voice, and video interviews on desktop web. Before you get started, a few things to note where Physical Intelligence is not yet supported:

• ❌ Mobile web usability studies

• ❌ Mobile app usability studies

• ❌ Video upload

• ❌ Live video analysis during the interview (e.g. flipping to analyze a camera feed in real time)

Can participants upload any image file type? Outset supports .png and .jpg image formats. Check with your Outset account team if you have questions about specific formats.

What happens if a participant submits the wrong image and re-upload is not enabled? The interview will continue with the submitted image. The AI will still analyze and summarize what it sees, and the image will be tagged accordingly.

Can I add more than one Participant Upload section to a guide? Yes — you can include multiple Participant Upload sections in a single guide. Each section allows one photo upload.

Is Physical Intelligence available for mobile web and mobile app usability studies? Not currently. Physical Intelligence is supported on desktop web for text, voice, and video interviews. Mobile web and mobile app usability and video upload are on the roadmap for a future release.

Hope this helps! If you have any further questions, please reach out to our team at support@outset.ai or via chat.