How I Automated 847 Fleet Maintenance Issues (And Never Missed Another Service Call)

The $8,500 Wake-Up Call

3:17am. My phone buzzes. It's dispatch.

"One of our trucks is stranded on the M1. Engine fault. The client's shipment won't make the morning delivery window."

I pulled up Fleetio on my phone, still half-asleep. There it was—a maintenance issue flagged 14 days ago for that exact truck. Oil pressure warning. Scheduled for service at our Brisbane depot. Never followed up on.

The cost that night:

• $2,100 emergency tow
• $3,800 urgent engine repair
• $2,600 penalty for missed delivery
• One very angry client (now an ex-client)

But the real damage? I had 118 other trucks with open maintenance issues sitting in Fleetio that I didn't even know about.

The Core Problem With Fleet Management Software

Fleetio is excellent. So is every other fleet management platform. They collect all the data you need—service schedules, fault codes, meter readings, inspection notes.

But here's the problem: data sitting in a dashboard doesn't prevent a truck from breaking down.

Someone has to:

• Check the dashboard daily
• Identify which issues are urgent
• Call the right service center
• Follow up if they don't respond
• Verify the work was done
• Update the system

With 120 trucks across 8 service centers, that's a full-time job. And the moment someone gets busy, issues slip through.

The Automation Framework That Fixed It

After that 3am call, I built a system that does all of this automatically. No human required unless something actually needs a decision.

Here's how it works:

1. Automated Issue Detection

Every morning at 6am, the system pulls all open maintenance issues from Fleetio via their API into a Google Sheet.

Not just issue titles—the full context:

• Which truck (ID, make, model, current location)
• What the issue is
• How long it's been open
• Which service center was assigned
• Last contact attempt (if any)

This Google Sheet becomes the "command center" where everything happens.

2. AI-Powered Prioritization

Raw data isn't useful. You need logic to determine what actually matters.

This is where Gemini AI comes in. A Google Apps Script runs every hour and feeds the open issues into Gemini with a simple prompt:

"Based on this maintenance data, which issues require immediate action? Consider: issue age, vehicle criticality, service center response patterns, and safety implications."

The AI returns a prioritized list with reasoning. For example:

It also identifies patterns humans miss:

• "Melbourne depot has 7 open issues over 21 days old—consider escalating to regional manager"
• "Truck #89 has had 3 electrical issues in 45 days—potential recurring problem worth investigating"

3. Automated Outreach Sequence

Once issues are prioritized, the system doesn't wait for a human to take action.

Day 0: Issue detected and logged

Day 2 (if no update in Fleetio): Automated email sent to the assigned service center:
"Truck #47 has an open brake warning issue for 32 days. Can you provide a status update or estimated completion date?"

Day 4 (if still no response): Automated phone call placed to the service center. The AI voice asks for an update, captures the response, and logs it.

Day 6 (if critical and still unresolved): Escalation email to my inbox with full context and suggested actions.

The key insight: Most service centers respond after the first email. They're not ignoring you—they're busy and forget to update systems. A simple automated nudge solves 80% of follow-ups.

4. Smart Data Sync Back to Fleetio

When a service center responds (via email or phone), the AI doesn't just file it away. It:

1. Extracts the key info ("completed yesterday" → actual completion date)
2. Cleans and formats the response
3. Updates the corresponding Fleetio issue via API
4. Closes the issue if complete, or reschedules follow-up if still pending

Everything stays synchronized. Fleetio remains the source of truth, but now it's actually up to date.

The Results (Real Numbers)

I've been running this system for 6 months. Here's what changed:

• 847 maintenance issues automatically processed
• Zero critical issues missed
• ~15 hours/week saved on manual follow-ups
• Average resolution time: Dropped from 18 days to 7 days
• Service center response rate: 94% (up from ~60% when I was calling manually)

More importantly: I haven't had another 3am emergency call about a preventable breakdown.

The Tech Stack (What You Actually Need)

This isn't a complex enterprise system. The entire automation runs on:

• Fleetio: Fleet management platform (you probably already have this)
• Google Sheets: Data hub and command center
• Google Apps Script: The glue that connects everything (free, built into Google Workspace)
• Gemini API: AI logic for prioritization and response parsing
• Email service: Automated outreach (can use Gmail API or any email service)
• Voice AI service: Automated calls for non-responsive service centers

Total monthly cost: ~$50-80 (mostly Gemini API calls and voice minutes)

Compare that to hiring someone to manually track 120 trucks, or the cost of one missed breakdown.

Why Most Fleet Managers Don't Do This

It's not because the tools don't exist. They do. Fleetio even has a comprehensive API built for exactly this kind of integration.

The problem is implementation. You need to:

• Understand how APIs work
• Build the Google Apps Script logic
• Design the AI prompts that actually return useful prioritizations
• Handle edge cases (what if the API times out? What if the service center email bounces?)
• Set up proper error logging so you know when something breaks

Most fleet managers don't have the time (or frankly, the interest) to learn this. They're busy managing fleets.

And that's fine. You can absolutely hire someone to build this for you. The ROI is immediate—the first prevented breakdown pays for the setup.

The One Thing You Should Automate First

If you're not ready to automate the entire workflow, start here:

Automated daily email digest of open issues > 14 days old.

That's it. Every morning, you get an email with:

• Which trucks have old open issues
• Which service centers aren't responding
• What needs your attention today

This alone would have prevented my $8,500 wake-up call.

The full automation comes later, but just getting visibility into what's falling through the cracks is 80% of the value.

Final Thoughts

Fleet management software is only as good as the humans using it. And humans miss things.

Automation doesn't replace good judgment—it replaces the tedious, repetitive tasks that prevent good judgment from being applied in the first place.

You shouldn't be manually checking dashboards, calling service centers, and updating spreadsheets. Your job is to make strategic decisions about your fleet.

Let the automation handle the rest.