Claim records from customer photos
A customer sends a photo, and AI fills in the claim
A CS team that gets claims as photos, in manufacturing, food, consumer goods or field installation, set up a flow where AI reads the product name, lot, best-before date and damage from each photo and fills in the claim record.
The problem
- 01
Typing details off a photo
Agents zoomed into photos to copy product names and lot numbers by hand, and the next inquiry had to wait.
- 02
Asking customers twice
Agents called back to confirm what the photo already showed, so customers explained everything twice.
- 03
No running count of claim types
Claims sat in chats and notes, so no one could tell which products kept having which problems.
How Trackit solves it
- 1
The customer sends a photo and a note through a web form
The customer submits a photo of the package or the site along with what happened. Submitting the form creates a new claim record, and the attached photo starts the workflow. Agents have nothing to copy first.
- 2
AI reads the product details and the problem type
AI file analysis reads the product name, lot number, best-before date and damage from the photo, and picks a problem type from the options the team set (damaged package, foreign object, spoiled, short count). A create-or-update-record block finds the customer by the phone number on the form, or creates one for a first-time customer.
- 3
The claim is filled in and handed to an owner
An update-record block writes the extracted values and the matched customer onto that claim. A create-task block assigns a review task to the claim owner, who gets notified. On a dashboard, claims grouped by product and type show which problems keep coming back.
- TriggerFile attachedA photo from the web form lands on a new claim record.
- AIAI file analysisReads product, lot, best-before date and type from the photo.
- ActionCreate or update recordFinds the customer by phone number, or creates one.
- ResultUpdate recordWrites the values and the customer onto the claim.
- ResultCreate taskAssigns a review task to the claim owner.
What changed
Before Trackit
With Trackit
More use cases
Phone sales teamCustomer records filled from calls and photos
Work-phone calls arrive through a CTI webhook and car photos through a web form, and AI reads both into one customer record.
B2B manufacturersCustomer requests from intake to resolution
Sample, defect and service requests are linked to the customer and deal, with team assignment, approval and outcome on one record.
Subscription and maintenance businessesEarly alerts for at-risk customers
AI regularly reads each customer’s activity and renewal date, scores churn risk, and alerts the owner about at-risk customers.