Type it in
The most direct way: in Records, click New record, choose the form (and the project, if you have several), then click Start draft. The record opens ready to fill in, and you can set its location on the record. Conditional rules show and hide fields as you go, so you only see what's relevant.
Take a photo
Add a photo to a photo field straight from your device. From a single picture, the agent can also draft a whole record, proposing values from what it sees for you to check and submit. Hidden location and device data is stripped from every photo on upload. If an admin turns on Capture photo location (Settings → Privacy & data), the photo's GPS position and capture time are read first and shown on the record.
In the chat, send the photos for one record together, in one message. The agent can only file photos from your most recent message with photos (in the mobile app, only those in the message you're sending), and only onto a photo field that can be filled in at the record's current step. On the web, chat photos aren't kept after you refresh the page, and in a long chat with many uploads only the latest are sent (the agent tells you when that starts). To use an older photo, send it again.
Describe it in a sentence
Tell the agent what happened ("crack in the east stairwell slab, looks structural, high severity") and it fills in the matching fields, finds the location, and shows you a draft to review. Nothing is submitted until you say so.
From the mobile field app
The mobile app is built for site work: create and fill in records on your phone and attach photos. With no signal, Started on site lets you write a note and take photos on the device, then finish and submit the record once you're back online. See the mobile field app.
Whichever way a record starts, it follows the same workflow and the same required-field rules.
Filling in records needs a seat that can create them. With a view-only seat, you can open and read records but not create them or use the agent.