The scene
It’s a Tuesday afternoon and Sandra has been dispatching for fourteen years. Three weeks ago, her fleet rolled out a new AI-assisted routing tool — the kind that’s supposed to cut deadhead miles and flag capacity issues before they become late-delivery problems.
She’s still running her own spreadsheet on the side.
Not because the tool is broken. Because two weeks in, she flagged that one of its suggested routes ran a driver straight through a weight-restricted bridge — and the response she got back was closer to «just trust the system» than «good catch, let’s look at that.» So now she quietly double-checks everything before it goes out, tells the owner things are «working fine,» and keeps her spreadsheet running in the background just in case.
Nobody lied. Nobody sabotaged anything. Sandra just learned, in one conversation, that flagging a problem wasn’t safe. So, she stopped flagging problems.
The pattern
Most owners read a story like Sandra’s as a people problem. She doesn’t want to change. She’s stuck in her ways. She doesn’t trust the tech.
It’s not that. It’s a psychological safety problem — and it’s the single biggest reason AI tools in small fleets get quietly abandoned or worked around, long before anyone admits the rollout failed.
A few things make this sharper in trucking than in a typical office rollout:
- Prior tech trauma. Most fleets have already lived through at least one bad software rollout — an ELD scramble, a TMS nobody was trained on. «We tried this before» isn’t resistance. It’s pattern recognition.
- A surveillance history. ELDs, GPS, dash cams — drivers and dispatchers already associate technology with being watched. Any new AI tool has to answer, «is this here to help me, or catch me?» before it can do anything else.
- Competence anxiety. Sandra’s value isn’t her typing speed — it’s fourteen years of judgment about routes, customers, and drivers. A poorly introduced AI tool can quietly say «the thing you’re good at doesn’t matter anymore,» even when that’s not the intent at all.
- The owner/driver gap. The owner usually feels safe — they chose the tool; they see the ROI case on paper. The dispatcher or driver who has to use it daily often feels like it happened to them. Safety has to be checked per role, not fleet wide.
The reframe
Here’s the shift that matters: psychological safety isn’t something your team has or doesn’t have. It’s something leadership builds — or fails to build — in the first few weeks of a rollout.
Which means low adoption usually isn’t a willingness problem in the way owners think. It’s an unaddressed safety question wearing a willingness costume. Someone hasn’t decided the tool is bad — they’ve decided it isn’t safe to be honest about what’s actually happening when they use it.
That distinction changes what you fix. You can’t train your way out of a safety problem. You can only build your way out of one — through how leadership responds the first few times someone raises a concern.
One thing you can check this week
Three signals your team doesn’t feel safe with a new tool yet:
- Silence when it comes up in meetings. Not agreement — silence.
- A sudden spike in «it’s working fine.» Real rollouts have friction. If nobody’s reporting any, they’ve stopped reporting.
- Workarounds reappearing. Paper logs, side spreadsheets, texting instead of using the system — people quietly reverting to what felt safe before.
One first move, before spending another dollar on tools: run a short, no-blame «what’s not working yet» conversation with the people actually using the system — separate from any performance discussion. The goal isn’t to fix everything in that conversation. It’s to prove that flagging a problem doesn’t cost anything. That’s the whole game in week one.
(For bilingual teams — this includes language safety. If your Spanish-speaking drivers or dispatchers are getting translated materials after the fact instead of from day one, that’s its own quiet signal about who the rollout was really built for.)
Where this fits
This is domain one of twelve in what I think of as the path to real AI culture adoption in small fleets — not the tech stack, the human side that determines whether the tech stack ever gets used the way it was meant to. Over the next eleven issues we’ll work through the rest: mindset, leadership, communication, resistance, governance, and more.
If you’re wondering where your own operation actually stands before spending anything on tools, that question — is this team safe enough to be honest with me right now — is usually the first thing I look at in an assessment. Happy to talk through it if it’d be useful.
YBOR Logistics AI
Helping logistics companies turn AI, automation, and operational data into better decisions, stronger systems, and more reliable fleet operations.
