Hitchyard LLC - AI Onboarding and Compliance for Freight Driver Discovery

Company Overview
Hitchyard is a Salt Lake City delivery service connecting drivers with warehouses and manufacturers. Currently, operations rely on phone calls, texts, spreadsheets, and manual checks. Hitchyard is building an AI-powered delivery system that automates onboarding, driver-to-load matching, live tracking, pricing forecasts, exception handling, and payments.

Project Description
Students on this project will build a driver and shipper verification and compliance component for the Hitchyard system that will automatically confirm that trucking companies are legal, insured, and safe. Students will also use our AI system to find and rank reliable drivers and onboard them into the system.

Key Technical Components

Salesforce: Stores onboarding data, driver performance scores, load tracking, and market history. Agents reference Salesforce for real-time decisions.

AI Engine: Dify + Groq (Llama models) for vetting, matchmaking, exception detection, and structured outputs like PDF or JSON.

User Interface: Mobile-First Web V0, immediately accessible by drivers via link, and serves as a blueprint for future Flutter or React Native apps.

Integrations: FMCSA (driver safety), Ansonia (KYB and business credit check), DAT (pricing and carrier ranking), Melio/Mercury (payments), Intuit/QuickBooks (accounting).

API Project Launch Sequence Students develop the APIs in this order:
1. Salesforce schema – stores operational data for all agents
2. Onboarding APIs – FMCSA, Ansonia KYB and business credit check
3. DAT & Matchmaking – pricing and carrier ranking logic
4. Fintech APIs – Melio, Mercury, Intuit/QuickBooks payments

Development Strategy

Mobile-First Web V0 / Functional Prototype Phase Focus on AI and logic: The same Dify agents and Salesforce logic work whether the interface is web or native app.

Rapid access and iteration: Drivers and shippers can use the system immediately via a link. Future app ready: Web pages provide wireframes, API connections, and functional flows for later Flutter/React Native builds.

Guardrailed AI (Groq + Llama) Goal: Zero token spend and instant processing while keeping operations safe.

Primary Model: Llama 3.1 8B via Groq – vetting, invoice generation, and monitoring exceptions

Secondary Model: Llama 3.3 70B via Groq – matchmaking, rate locking, and contract template drafting

Guardrail: AI provides guidance but Salesforce makes the final decision, for example, it confirms driver insurance is valid before completing a load assignment

Student Takeaways

Students will gain experience building a real-world delivery platform by:

Integrating multiple APIs in sequence
Designing Salesforce data models
Automating financial workflows
Deploying AI agents
Building a mobile-first interface ready for future native apps