AI Assist Workflow
AI Assist is Benchmrk's AI-powered feature for finding relevant survey marks near a location. This page walks through the complete workflow from activation to interpreting results.
Overview
The AI Assist workflow has four stages:
- Activate pin-drop mode — Enter the AI analysis interface.
- Place your pin — Click the map to set your location of interest.
- Configure preferences — Tell the AI what kind of marks you need.
- Review results — Interpret the AI's recommendations.
Stage 1: Activate pin-drop mode
To start an AI analysis:
- Click the AI Assist button in the sidebar, or use the AI mode toggle.
- The interface enters pin-drop mode — the map cursor changes to a crosshair.
- The sidebar shows the AI Assist panel with instructions.
AI Assist requires an active subscription with available usage quota. If your quota is exhausted, you'll see a message prompting you to wait for the monthly reset or upgrade your tier.
Stage 2: Place your pin
With pin-drop mode active:
- Click anywhere on the map to drop a pin at your location of interest.
- The pin appears on the map at the clicked coordinates.
- The sidebar updates to show the preferences form.
The pin location tells the AI where to focus its search. It analyses marks within a radius of this point.
Repositioning the pin
If you placed the pin in the wrong spot, simply click a new location on the map. The previous pin is replaced with the new one.
Stage 3: Configure preferences
After placing your pin, the preferences panel lets you fine-tune the analysis. Each preference affects how the AI ranks and filters nearby marks.
| Preference | What it does |
|---|---|
| Project type | Tells the AI what kind of work you're doing (e.g. construction, cadastral survey, engineering) so it can prioritise relevant mark types |
| Priority | Sets whether you value proximity, accuracy, or a balance of both |
| Max distance | Limits how far from the pin the AI should search |
| Mark status | Filters by mark condition (e.g. only active marks, or include destroyed marks) |
| Time window | Restricts results to marks surveyed within a time period |
| Horizontal class minimum | Sets the minimum horizontal accuracy class |
| Vertical class minimum | Sets the minimum vertical accuracy class |
| Additional notes | Free-text field for any extra context the AI should consider |
For detailed explanations of each preference, see AI Preferences.
Once you've configured your preferences, click Analyse to start the AI analysis.
Stage 4: Review results
The AI processes your request and returns a ranked list of recommended survey marks. Results include:
- Ranked recommendations — Marks ordered by relevance to your criteria.
- Reasoning — For each recommended mark, the AI explains why it was selected.
- Mark details — Each result links to the full point details panel so you can inspect the mark's metadata.
- Distance — How far each recommended mark is from your pin location.
Interpreting the results
The AI considers multiple factors when ranking marks:
- Proximity to your pin location.
- Mark status — Active marks are generally preferred over destroyed ones.
- Accuracy class — Marks meeting your horizontal/vertical class minimums rank higher.
- Relevance to project type — The AI weighs mark characteristics against your stated project needs.
- Recency — More recently surveyed marks may be preferred depending on your time window setting.
The AI's recommendations are a starting point. Always verify mark details in the point details panel before using them for professional survey work.
Quota consumption
Each AI analysis consumes one unit from your monthly usage quota. The quota counter updates immediately after the analysis completes. See AI Quotas for details on quota limits per subscription tier.
What happens next
After reviewing results, you can: