AI-Improved Category Search and Photo Analysis helps residents and staff find the correct request category faster and more accurately when submitting new requests. By using AI to understand meaning and visual context, these capabilities reduce friction during submission and help ensure requests are routed correctly the first time.
This shortens the submission time for residents and staff and helps reduce follow-up work caused by miscategorized requests, supporting faster resolution overall.
This functionality includes two AI-assisted features that work together to improve category discovery during request submission.
AI-Improved Category Search
When a user searches for a category, the system uses AI to evaluate the meaning of the search term and compare it against the category’s title.
Instead of relying only on exact matches, the system looks at conceptual similarity. This allows relevant categories to appear even when the wording used by the resident or staff member does not exactly match the category name or keywords.
For example, a search for “Sparkler” may surface an “Illegal Fireworks” category even if the word “Sparkler” is not explicitly listed. This helps users find the right category more quickly and reduces guesswork during submission.
This behavior applies to:
Resident request submission flows
The CRM New Request form used by staff when entering requests on behalf of residents
AI-Assisted Photo Analysis
When a photo is included with a request, AI can analyze the image to help identify the most relevant category.
At a high level:
The image is analyzed to identify descriptive labels
Those labels are converted into embeddings
The embeddings are compared to category embeddings to determine the closest match
This allows the system to suggest relevant categories even if the user does not enter a search term.
Important Notes
Only the first uploaded image is analyzed
Photo analysis runs only when the photo is uploaded before category selection. This flow is the default in Marketplace mobile apps and is available in embedded portals when explicitly enabled.
About Embeddings
Embeddings are a way for AI to represent words, phrases, and concepts based on meaning rather than exact text. In SeeClickFix 311 CRM, embeddings are generated from:
Category titles
Labels generated during photo analysis
These embeddings allow the system to compare concepts and determine relevance, improving category discovery without changing how categories are configured or managed.
The system does not automatically assign categories. It provides suggestions to help users make the correct selection.
Organization Setting
AI-Improved Category Search and Photo Analysis are controlled by a single organization-level setting.
When enabled:
AI-Improved Category Search is used during category search
Photo analysis is used when a supported request flows upload photos before category selection
When disabled:
Category discovery uses exact matching behavior
Keywords continue to function normally
Unsupported Request Flows
AI-assisted category features are not supported in the SeeClickFix 311 CRM container app.
The container app allows users to submit requests across multiple organizations. Because the organization is not known until the user selects a location and category, organization-level AI settings cannot be applied. As a result, neither AI-Improved Category Search nor AI-Assisted Photo Analysis is available in the container app.