Skip to content
AtheronLABS

You're visiting from the United States. Prices are shown in US dollars. Not right?

labs@atheron:~/industries/trucking/ai$ train agent --grounded

Fleet safety assistant

Answers from your own safety manual, kept private.

A fleet's safety manual is long, and the same questions come up every week. An assistant that answers from your current manual and cites the section saves your safety staff time, and an incident report filed from the cab keeps the details while they are fresh.

// what matters here

What is different in this industry.

01

Your manual, cited

Every answer names the policy and section it comes from, so staff can check before acting.

02

Private by design

An open-source model in your own cloud account keeps driver and incident records out of anyone else's systems.

03

Reports while they are fresh

Photos, location and the ELD data around the event are attached when the driver files the report.

04

No safety decisions

The assistant helps people find answers. Decisions about drivers and incidents stay with your safety manager.

// example projects

Priced examples for this service.

Each one is hypothetical, labelled as such, and priced live from our rate card. Open any of them in the estimator and make it yours.

Example projectWeb appMobile appAIOpen-source model

Safety and incident assistant

Hypothetical. Not a client, not a result.

// The problem

Drivers report incidents by phone hours later, details are lost, and staff dig through a thick safety manual and past reports to answer the same questions.

// What we would build

An incident report in the driver app with photos, location and the ELD data around the time, and an assistant that answers staff and driver questions from your own safety manual and policies, citing the section. It runs on an open-source model in your own cloud account so driver and incident records stay private.

// What is in it

  • Incident and near-miss reports from the cab, with photos and location
  • ELD data around the time of the event attached to the report
  • An assistant that answers from your safety manual and cites the section
  • An open-source model on GPUs in your own cloud account
  • An evaluation set of real questions, run before every change

// Stack

  • Llama or Mistral (open-source)
  • vLLM
  • PostgreSQL with pgvector
  • React Native
  • AWS GPU

// estimate

Build
≈ US$160,000 to US$244,000, delivered within 29 weeksCAD 227,600 to 347,900
Hosting
≈ US$7,020 a monthCAD 10,000 a month
Support
≈ US$2,500 a monthCAD 3,565 a month

Prices in your currency are estimates from today's Bank of Canada rate. All invoicing is in CAD or USD.

AI route
Open-source model
Model running cost
≈ US$6,530 a monthCAD 9,300 a month

Timeline by milestone

Discovery
2.9 to 3.4 weeks
Specification and evaluation plan
0.4 weeks
Design approved
3.9 to 6.4 weeks
Core features
1.4 to 1.9 weeks
Full build
0.9 to 1.4 weeks
First build on devices
1.9 to 3.4 weeks
Working pilot
4.4 to 7.4 weeks
Testing and fixes
0.9 weeks
Store submission
0.9 to 1.9 weeks
Launch
0.9 to 1.4 weeks
Production
0.9 to 1.4 weeks

Outside our hands, and added to the calendar

App store review
1 week
API access from your ELD provider
1 to 4 weeks

How it is paid

Deposit 20%
CAD 45,460 to 69,520
Discovery 4.1%
CAD 9,319.30 to 14,251.60
Specification and evaluation plan 7.7%
CAD 17,502.10 to 26,765.20
Design approved 4.1%
CAD 9,319.30 to 14,251.60
Core features 3.4%
CAD 7,728.20 to 11,818.40
Full build 2.2%
CAD 5,000.60 to 7,647.20
First build on devices 9%
CAD 20,457 to 31,284
Working pilot 23.1%
CAD 52,506.30 to 80,295.60
Testing and fixes 1.1%
CAD 2,500.30 to 3,823.60
Store submission 3%
CAD 6,819 to 10,428
Launch 4.1%
CAD 9,319.30 to 14,251.60
Production 8.2%
CAD 18,638.60 to 28,503.20
Holdback, 30 days after launch (10%)
CAD 22,730 to 34,760
GPU time for training and testing, at cost
CAD 300. Billed up front, at cost, outside the milestones

// questions

Questions we get about this.

Can the safety assistant decide whether an incident was preventable?

No. It helps staff find the policy and past reports. That decision stays with your safety manager.

What happens when our safety manual changes?

The new version replaces the old one in the index, and the evaluation set is run again before the change goes live.

Why an open-source model for fleet safety records?

Driver and incident records are personal. Running the model in your own cloud account keeps them there. A frontier model costs less to start if you are comfortable sending them to a provider.

// next

Not quite your project?

Tell us what you have in mind. We will come back to you with a range and the questions that would narrow it. Or book a call and talk it through.

Private AI safety assistant and incident reporting for fleets | Atheron Network Labs