Skip to main content

I’m a Solver. How does duplicate detection work?

Learn how to reduce duplicate reports.

Updated over a week ago

Duplicate Detection Overview

When a Snapper selects an incident type, we identify the Solver it will be sent to. If that Solver has the Duplicates feature enabled, we’ll automatically check for potential matches and show them to the Snapper.

If the Snapper confirms a match, no new report is created—this is recorded as a deflection, helping reduce duplicate reports.

How are matches calculated?

We use AI to examine a range of factors when considering a potential match including:

  • Image recognition

  • Location proximity

  • Incident type

  • Date reported

  • Incident description

If any potential matches are found (with an 85% confidence level or higher) up to 3 matches will be shown to the Snapper who can then choose to match their current report to one of these potentials.

What happens when a match is confirmed?

When the Snapper matches their would-be report to an existing one, they exit the reporting flow and a report deflection is recorded.

An additional report is not created by this Snapper.

You will be notified the first time an existing report is matched and results in a deflection.

Where can I see which reports have been matched?

If you have the Duplicates feature enabled, you will see a “Deflections” column in the report list showing the number of times a Snapper started a report but chose an existing match instead of submitting a new one. This information will also be included on the report details page.

If you don’t have the duplicates feature enabled, you will see a “Potential Duplicates” column the report list showing the number of other reports created within 25 meters and 14 days of the current report.

What kind of reports cannot be returned as a match?

We won’t show reports for potential matches that:

  • Are flagged as sensitive by our AI content moderation.

  • Do not contain a photo.

  • Have been cancelled.

What information is shown on matched reports?

When potential duplicates are shown to a Snapper, each matched report includes:

  • Time the incident was reported

  • Proximity (in meters) from the Snapper’s current report

  • Report photo(s)

  • Map view of the incident location

  • Incident type

  • Description of the issue

These details are provided to help the Snapper confidently identify if their report matches an existing one.

Did this answer your question?