As a daily commuter using Commute.ai, I want to be able to input additional requirements or preferences through text or voice after seeing initial route suggestions, so that I can quickly filter and adjust the routes to match my specific needs without manually checking each option.
Acceptance Criteria
Given: User has searched for a route from Point A to Point B
The app displays initial route suggestions (e.g., Kamppi to Exactum)
A prominent text input field or voice button is visible below the route results
The input field has placeholder text like "Refine your routes... (e.g., 'only show routes through Kalasatama')"
When: User enters a refinement prompt
User types or speaks a natural language request (e.g., "Only show routes that go through Kalasatama")
The AI agent processes the request and understands the constraint
The system filters existing routes and/or fetches new routes that match the criteria
Then: Updated routes are displayed
Routes that don't match the criteria are removed or hidden
New qualifying routes are added if available
A visual indicator shows the active filter (e.g., chip/tag showing "Via Kalasatama")
User can see how many routes were filtered out (e.g., "Showing 3 of 8 routes")
User can clear the filter to return to original suggestions
Additional Scenarios
Scenario 1: Multiple Refinements
User can add multiple refinement prompts sequentially
Each refinement further narrows the results
All active filters are visible and individually removable
Scenario 2: Voice Input
User can tap a microphone icon to speak their refinement
Speech-to-text converts the input
Same filtering logic applies as text input
Scenario 3: No Matching Routes
If no routes match the criteria, show helpful message
Suggest relaxing constraints or trying alternative phrasing
Offer to show "closest matches" that partially satisfy the request
Example Refinement Prompts
- "Only routes with less than one transfer"
- "Avoid metro, prefer trams"
- "Routes that pass by a coffee shop"
- "Most scenic option"
- "Least crowded at this time"
- "Show me the route with most outdoor walking"
Technical Considerations
AI agent must parse natural language and extract route constraints
Integration with HSL API to fetch additional routes if needed
Real-time filtering of existing route data
Support for both text and voice input modalities
As a daily commuter using Commute.ai, I want to be able to input additional requirements or preferences through text or voice after seeing initial route suggestions, so that I can quickly filter and adjust the routes to match my specific needs without manually checking each option.
Acceptance Criteria
Given: User has searched for a route from Point A to Point B
The app displays initial route suggestions (e.g., Kamppi to Exactum)
A prominent text input field or voice button is visible below the route results
The input field has placeholder text like "Refine your routes... (e.g., 'only show routes through Kalasatama')"
When: User enters a refinement prompt
User types or speaks a natural language request (e.g., "Only show routes that go through Kalasatama")
The AI agent processes the request and understands the constraint
The system filters existing routes and/or fetches new routes that match the criteria
Then: Updated routes are displayed
Routes that don't match the criteria are removed or hidden
New qualifying routes are added if available
A visual indicator shows the active filter (e.g., chip/tag showing "Via Kalasatama")
User can see how many routes were filtered out (e.g., "Showing 3 of 8 routes")
User can clear the filter to return to original suggestions
Additional Scenarios
Scenario 1: Multiple Refinements
User can add multiple refinement prompts sequentially
Each refinement further narrows the results
All active filters are visible and individually removable
Scenario 2: Voice Input
User can tap a microphone icon to speak their refinement
Speech-to-text converts the input
Same filtering logic applies as text input
Scenario 3: No Matching Routes
If no routes match the criteria, show helpful message
Suggest relaxing constraints or trying alternative phrasing
Offer to show "closest matches" that partially satisfy the request
Example Refinement Prompts
Technical Considerations
AI agent must parse natural language and extract route constraints
Integration with HSL API to fetch additional routes if needed
Real-time filtering of existing route data
Support for both text and voice input modalities