When you enable AI Dispatcher and mark a job for automatic assignment, the system analyzes hundreds of variables in milliseconds to find the best technician match. This isn't calendar Tetris—it's solving a multi-constraint optimization problem that balances skills, location, availability, workload, and route efficiency simultaneously.

Here's what happens under the hood.

Before optimization starts, AI Dispatcher filters out technicians who can't physically do the job. Every assignment must pass these hard constraints:

Fail any constraint? That technician is out. No amount of optimization overrides these—they ensure compliance and prevent impossible assignments.

Example: A commercial HVAC repair in downtown requires EPA 608 certification and a lift-equipped van. Only technicians with both pass this filter.

Now the engine scores every eligible technician across five optimization factors:

The system calculates real driving time from the technician's current location (or last job) to the new site. It uses road networks and traffic patterns, not straight-line distance.

Example: Sarah is 5 miles away via highway. Tom is 3 miles through downtown. Sarah scores higher—her route takes 8 minutes vs. Tom's 18.

Meeting the minimum isn't the same as being the best fit. AI Dispatcher ranks technicians by expertise level, not just yes/no qualification.

Example: Job requires "HVAC Certified." Both Mike (basic cert, 6 months experience) and Lisa (master cert, 8 years) qualify. Lisa scores higher for a complex chiller repair. Mike might score higher for a routine filter change.

The engine looks at each technician's schedule utilization over the next 2-4 weeks, not just today. Goal: keep everyone between 70-85% capacity—busy enough to be efficient, not so slammed that quality drops.

Example: Jake is at 45% utilization this week, Emma at 82%. All else equal, Jake gets the job to balance the team.

If this client has service history, the system checks who's worked there before. Returning the same technician earns a continuity bonus—customers like familiar faces, and the tech already knows the equipment.

Example: This is the third service call for Murphy's Restaurant. Carlos did the previous two and knows their walk-in cooler quirks. He gets a scoring boost.

This is the sophisticated part. AI Dispatcher doesn't just look at one job—it simulates how this assignment affects the technician's entire day.

Example: Amy has jobs at 9am (north) and 3pm (south). The new 11am job is either downtown (creates a linear north→center→south route) or west side (forces backtracking). Downtown job scores higher for route efficiency.

Every assignment gets a confidence score (0-100%) showing how optimal it is:

Example: AI assigns Tom with 68% confidence: "✓ Skills match perfectly, ✓ Customer continuity, ⚠ 45-minute drive, ⚠ Workload at 87%." You see the trade-offs and can override if needed.

AI Dispatcher monitors for changes and re-runs optimization when things shift:

Example: Carlos finishes his 10am job 30 minutes early. He's now the closest tech to a 1pm appointment currently assigned to Maria. AI suggests reassigning to Carlos—but only suggests, never auto-swaps. You keep control.

AI Dispatcher adjusts over time based on actual outcomes:

The AI doesn't replace your judgment—it amplifies it. You set the rules (skills, areas, limits). You review questionable assignments. The AI does the computational heavy lifting: testing hundreds of combinations per second to find matches you'd never calculate manually.

Every assignment shows why that technician was chosen, not just who:

"✓ Skills match perfectly
✓ Only 8 minutes away
✓ Route efficiency: 94%
⚠ Workload reaching 85%"

This transparency lets you override when you have context the AI doesn't: maybe this technician prefers working with this client type, or you know they need lighter days this week for personal reasons.

The goal isn't full automation. It's making you 10x faster at dispatch while preserving the judgment calls that require human insight.