Why Your Engineers Are Doing Admin Work
I spent seven years as a mechanical engineer at FMG and MinRes before starting my AI consulting practice. And here’s something I noticed across both sites: the best engineers I worked with spent maybe 60% of their time actually engineering.
The rest? Admin work. Copy-pasting data between systems. Chasing approvals. Updating spreadsheets. Writing the same email for the fifth time that week.
Nobody talks about this. We just accept it as part of the job.
But when I started doing AI opportunity assessments for engineering and construction firms across Australia, I found something interesting. That 40% admin burden isn’t just annoying. It’s expensive. Really expensive.
The Actual Cost of “Just Part of the Job”
Let’s do the math on a typical 50-person engineering consultancy.
Say you’ve got 30 engineers averaging $120,000 per year in salary (conservative estimate). If they’re spending 40% of their time on administrative tasks instead of billable engineering work, that’s roughly $1.4 million in salary going to work that doesn’t generate revenue.
But it gets worse. Those same engineers could be billing at $150-200 per hour for actual engineering work. At 40% non-billable time, you’re looking at around $2.5 million in lost revenue potential per year.
For a 50-person firm, that’s the difference between steady growth and struggling to hit targets.
And here’s the thing: most directors I talk to don’t realise how bad it is. They see the symptoms (engineers working late, projects running over, quality issues) but they don’t connect it back to workflow inefficiency.
What This Actually Looks Like
I recently completed an assessment with a 120-person electrical engineering consultancy in Perth. When I interviewed their senior engineers, here’s what I heard:
“I spend two hours every Monday updating our project tracker from emails and conversations I had last week.”
“We have three different systems for managing drawings, and none of them talk to each other.”
“I’ve written the same tender response section probably 50 times. It’s always the same information, just reformatted.”
Sound familiar?
The engineers weren’t complaining. They’d just accepted it as normal. But when we mapped out where their time actually went, the leadership team was shocked.
Their most experienced senior engineer, billing at $185/hour, was spending 12 hours per week on administrative tasks. That’s $115,000 per year of his salary going to work that could be automated.
Engineering and construction companies are good at engineering and construction. That’s what they focus on. That’s what they hire for.
But most firms grow by adding people, not by optimising workflows. You need more capacity, so you hire another engineer. Then you hire another. Eventually, you’ve got layers of coordination overhead that nobody planned for.
And because engineers are problem-solvers by nature, they just figure it out. They build workarounds. They stay late. They make it work.
But that doesn’t mean it’s efficient.
The other factor: engineering firms are conservative about technology. And fair enough. You can’t risk project delivery on some flashy new tool that might break. So most firms stick with what works, even if what works is inefficient.
What AI Actually Does (Without the Hype)
I’m not going to tell you AI is magic. It’s not.
But here’s what AI is genuinely good at: pattern recognition and repetitive tasks.
If your engineers are doing something repetitive (writing similar reports, extracting data from documents, formatting information), that’s a pattern. And AI can learn patterns.
Take that senior engineer spending 12 hours per week on admin. Half of that was updating project documentation from meeting notes and email threads. We built him a simple AI tool using Microsoft 365 that automatically extracts key information and populates the right fields.
Now it takes him 2 hours per week instead of 6. That’s 4 hours back for actual engineering work. Over a year, that’s roughly $37,000 in recovered billable time from one person.
And it didn’t require some fancy custom system. We used tools they already had.
The Three Places AI Works Best in Engineering Firms
Based on assessments I’ve done across mining, infrastructure, and construction firms, here’s where I consistently see the biggest returns:
1. Document processing and data extraction
Engineers deal with massive amounts of documentation. Specifications, drawings, tender responses, compliance reports. AI can read these documents, extract relevant information, and populate your systems automatically.
No more copy-pasting. No more manual data entry.
2. Communication and coordination overhead
Email summaries. Meeting notes. Status updates. Project reporting. These tasks follow patterns, and AI can handle a lot of the grunt work.
Your engineers can review and approve instead of writing from scratch.
3. Knowledge retrieval and reuse
Every engineering firm has decades of institutional knowledge buried in old projects, reports, and documentation. When someone needs to reference a similar project from three years ago, they either spend hours searching or just start from scratch.
AI can search your entire knowledge base and surface relevant information in seconds.
What This Means for Your Bottom Line
Let’s go back to that 50-person consultancy spending $1.4 million on administrative work.
If you could automate even 25% of that admin burden, you’d recover roughly $350,000 in salary costs. But more importantly, you’d free up capacity equivalent to hiring 3-4 new engineers without actually hiring anyone.
That’s real capacity growth without the recruitment costs, onboarding time, or overhead.
And for most firms, the implementation cost is far less than hiring even one additional engineer.
Why Most Engineering Firms Haven’t Done This Yet
Honestly? Because it sounds like hype.
Every second company is claiming they do AI now. Every software vendor has suddenly added “AI-powered” to their marketing. Engineering directors are getting bombarded with pitches from people who’ve never set foot on a construction site.
So the default response is scepticism. And that’s fair.
But here’s the difference: I’m not selling you AI. I’m helping you identify where your workflows are broken and whether AI happens to be the right solution.
Sometimes it is. Sometimes it’s just better processes. Sometimes it’s different software.
The assessment comes first. The technology comes second.
What Actually Works
If you’re thinking about this for your firm, here’s my honest advice:
Don’t start with the technology. Start with the workflows.
Map out where your engineers are actually spending their time. Talk to your project managers and senior engineers. Find the repetitive tasks that frustrate everyone.
Then look at whether AI can actually solve those specific problems. Not AI in general. Specific tools for specific tasks.
And start small. Pick one workflow. Prove it works. Then expand.
The firms that get real value from this aren’t the ones chasing the latest technology. They’re the ones who understand their own operations well enough to know exactly what needs fixing.
I help engineering and construction firms across Australia identify these workflow inefficiencies and implement practical AI solutions that actually deliver ROI. No hype, no overselling, just honest assessments from someone who’s been on the tools.
If you’re wondering whether your firm could benefit from this kind of analysis, let’s have a conversation. I offer a no-obligation AI opportunity assessment where we map out your workflows and identify specific areas where you’re losing time and money.
You might find it’s worth millions. Or you might find you’re already running efficiently. Either way, you’ll know for certain.