Skip to content

For NEMT dispatch teams

AI dispatch assistant for NEMT operators

On the dispatch software you already run.

Your dispatchers ask in plain English, such as which pickups are running late or who is free near an address at 2 p.m., and get answers from your own live trip data. It reads only until you say otherwise, a person approves anything that would change a trip, and every answer is logged. Start with the $950 AI Workflow Audit; the build is a fixed price from $4,500.

Book the $950 audit

What your dispatchers can ask

  • Which pickups are running late right now, and who is driving them?
  • Who is free near an address at 2 p.m., in a vehicle that fits the trip?
  • Which of tomorrow's trips still have no driver?
  • What did we bill a broker this week?
  • Which drivers' licenses or vehicle inspections expire this month?

Each answer lists the trips it came from, so a dispatcher can check it in seconds.

It works on top of your current software

The assistant reads your trip data where it already lives. During the audit we confirm how: your dispatch software's API, a read-only copy of its database, or regular exports. Whether your vendor allows that access is the first thing we check.

If we can't get read access to your trip data during the audit, you get the $950 back.

Where your data goes

  • Your trip data stays in your systems. The assistant has no copy of your database. It calls read-only tools when a dispatcher asks something.
  • A model receives only what a question needs. We use models from Anthropic or OpenAI, never use your data to train models, and turn on zero data retention where the provider offers it.
  • Each dispatcher sees only what they could already see. Tools run with the signed-in person's permissions.
  • Every question, tool and answer is logged, so you can show a broker or auditor exactly what happened.

We don't claim HIPAA compliance. If your broker contracts set rules for patient data, tell us at the start and we write how we'll meet them into the proposal.

Changes only after a person approves

The first version only reads. When you want it to act, such as drafting a driver assignment, the change waits in the assistant until a dispatcher approves it, and the approval is logged.

Who answers during your dispatch hours

Before the build starts, we agree a support window in your time zone and name the person who answers in it. You also get a short written update every working day.

How it starts

  1. AI Workflow Audit, $950, two weeks. We map your dispatch workflow, confirm data access and give you a build plan and a fixed quote. The $950 is credited if you build within 60 days.
  2. Build, fixed price from $4,500. A read-only assistant first, usually live in three to six weeks, with the code in your repository.
  3. Actions later, behind approval. Add drafted changes when your team trusts the answers.

We built a NEMT dispatch platform with an AI assistant for a client, with ten read-only tools over live trip data. That build is where this assistant comes from.

Screenshots

  • Mock-up on sample data: the dispatch assistant beside an existing dispatch board, answering which trips are late and who is free near a pickup, with a drafted assignment waiting for a dispatcher's approval.
  • Mock-up on sample data: the AI activity log, listing each question, the read-only tools used, the answer and who approved or declined each drafted change, ready to export for a broker or auditor.
Work

Related work.

Builds from our portfolio that use the same pieces.

Read next

Keep reading.

FAQs

Questions about cost, timing and ownership.

Cost, timing, ownership and your data.

  • No. It sits on top of the software you already use and reads its trip data. Your dispatchers keep working where they work now and ask the assistant when they need an answer.

  • Checking access is the first step of the audit. If we can't get read access to your trip data through an API, a read-only database copy or regular exports, you get the $950 back.

  • Not on its own. The first version only reads. If you want it to draft changes, each one waits for a dispatcher to approve it, and every approval is logged.

  • We don't claim HIPAA compliance, and we don't hold a healthcare compliance certification. The assistant reads through tools limited to what the signed-in person may see, a model receives only what a question needs, we never use your data to train models, and every AI action is logged. If your contracts set rules for patient data, tell us at the start and we write how we'll meet them into the proposal.

  • The audit takes two weeks. A read-only assistant is usually live three to six weeks after you approve the fixed quote.

Contact

Tell us what you want to build.

A few lines is enough. We'll reply within one business day with next steps or a time to talk.

We use these details only to reply to you. See our privacy policy.

We only use the cookies this site needs to work, such as keeping forms secure. Cookie policy