Multilingual AI Vision for Photo Analysis

Support Team
May 21, 2026
6 Views

Multilingual AI Vision for Photo Analysis

Laksana's AI can analyze photos sent via WhatsApp to verify and enrich work reports.

How It Works

  1. Staff sends a photo via WhatsApp
  2. AI analyzes the image content
  3. Identifies what the photo shows (e.g., clean floor, broken pipe, fire extinguisher)
  4. Cross-references with the text message for consistency
  5. Enriches the record with AI-generated description

Supported Languages

AI vision works in 8 languages:

  • English
  • Malay (Bahasa Malaysia)
  • Indonesian (Bahasa Indonesia)
  • Thai
  • Chinese (Simplified)
  • Vietnamese
  • Tagalog (Filipino)
  • Spanish

The AI description respects your organization's configured AI language setting.

Reference Photo Matching

  • AI compares field photos against historical benchmark photos
  • Improves accuracy over time as more reference photos are added
  • Helps verify that the correct area or asset was photographed
  • Detects significant changes from baseline conditions

EXIF Metadata

  • Extracts and preserves EXIF data from photos
  • Uses EXIF timestamp and GPS when available
  • Falls back to record context when metadata is missing
  • Metadata source is tracked for audit purposes

Use Cases

Work Verification

  • Confirms cleaning was actually done
  • Verifies maintenance was completed
  • Checks that the right area was photographed

Issue Enrichment

  • Identifies the type of damage or defect
  • Suggests severity based on visual evidence
  • Links to similar historical issues

Compliance Evidence

  • Verifies inspection photos show the correct checklist items
  • Confirms safety equipment is visible and in place
  • Flags missing or unclear evidence

Tips

  • Take clear, well-lit photos for best AI analysis
  • Include context in the photo (show the surrounding area)
  • Send multiple angles for complex issues
  • The AI works best with standard operational photos

Was this article helpful?

Help us improve our content by giving us feedback.

Related Articles