Hardest Part

The make-or-break challenge is building a real-time constraint solver that instantly replans entire fleet schedules around sudden disruptions without violating hard hour-of-service limits or cascading delays across the shift.

Min Viable Scope

The v1 focuses exclusively on single-depot regional delivery fleets with under 50 vehicles, executing only same-day dynamic rerouting based on traffic and driver availability. Deliberately exclude multi-depot load balancing, long-haul overnight compliance, and customer-facing SMS notifications.

Cold Start Problem

The predictive model requires baseline task duration data to enforce SLAs, but new customers rarely maintain clean logs. Break this by building a dedicated ingest pipeline that pulls 90 days of raw GPS telematics from existing fleet hardware to train the initial heuristics.

Time To First Value

2 weeks of onboarding to ingest historical telematics and calibrate routing heuristics

Data Moat Available

true

Technical Difficulty

High

Wedge

Begin by targeting emergency HVAC repair calls after hours and on weekends. This niche eliminates the need to replace daytime human dispatchers immediately while capturing high-margin missed revenue for the business. Once the system proves reliability in off-hours scheduling, expand into daytime overflow routing and eventually take over primary dispatch boards.

Timing

Voice and SMS-native LLMs now reliably handle multi-turn scheduling constraints and address parsing, while API-first field service management tools enable read and write access to schedule boards.

Why This ICP

Mid-sized trades operating 20 to 100 trucks feel acute margin pressure from dispatcher headcount and possess enough daily ticket volume to require dynamic routing, yet lack the budget to build custom in-house logistics software.

Size Of Prize

There are roughly 105,000 mid-sized home service businesses in the US, each representing a potential $30,000 annual spend to replace partial dispatcher labor. Multiplying 105,000 businesses by $30,000 yields an addressable prize of $3.15 billion.

Gap Narrative

Mid-sized field service operators rely on human dispatchers who manually match technician skills, truck inventory, and routing to incoming customer calls, causing schedule gaps and delayed response times. Existing software provides static routing algorithms but cannot dynamically negotiate appointment windows with customers via text or ingest live technician delays. The gap is an autonomous agent that handles the entire loop: ingesting the customer request, verifying parts, routing the tech, and adjusting the board in real-time.

Defensibility

The system builds workflow lock-in as it becomes the sole interface between the business and the technician schedule. However, core routing and LLM scheduling capabilities are largely commoditized. True defensibility only emerges if the agent builds a proprietary dataset of localized technician completion times, though it remains vulnerable to native AI features introduced directly by incumbent platforms.

Why This Thesis

An autonomous agent is required because dispatching is fundamentally a continuous state-space problem requiring live interaction with customers and techs, not just a static software dashboard. The agent acts directly on the environment, reading telemetry and writing schedule updates.

SAM

~$1B-2B targeting mid-market residential and commercial trades with 20+ trucks

SOM

~$50M-150M

TAM

~250k North American and European field service providers × ~$20k/yr ≈ ~$5B

Growth Rate

~12-18%/yr, driven by field labor shortages and rising fleet operating costs forcing higher daily route density

Paid Comparable Spend

~$40k-80k/yr per dispatch desk on human routing labor and legacy job management software modules

Kill Thresholds

• Human override rate exceeds 30 percent after 14 days of deployment
• Implementation and API integration time takes longer than 21 days
• D30 active usage by the primary dispatcher falls below 4 days per week
• Pilot conversion rate to paid contracts is under 20 percent after 90 days

Leading Metrics

• Percentage of daily jobs auto-assigned without human override
• Time spent by human dispatchers reviewing daily schedules in minutes
• Average travel time between job sites per technician
• Number of completed jobs per truck per day
• Time-to-first-value measured in days from pilot kickoff to first auto-dispatched route

What Proves Right

Mid-market fleets with 20 or more trucks deploy the AI Dispatch Controller and accept the automated routing suggestions without human intervention for 80 percent of daily jobs. Dispatchers reduce their daily schedule generation time from four hours to under twenty minutes. Customers sign $20,000 annual contracts after a 30-day pilot because the system demonstrably increases completed jobs per truck per day by at least one full job.

What Proves Wrong

Dispatchers manually override more than 40 percent of the generated routes because the system fails to account for uncaptured field constraints like specific technician certifications or truck inventory. Integrating with legacy systems requires custom engineering for each deployment, capping deployment velocity. Operations managers churn before day 60 because they refuse to trust automated routing logic over tenured human dispatchers.