I spent years in construction before moving into digital work. On a job site, I learned to see systems: materials flow in, work happens in sequence, handoffs need to be clean, or everything backs up.
I created workflows for almost every aspect of my job, from physical work—production-style component fabrication and preassembly—to back-office tasks, like quoting, contract creation, and project expense tracking.
Even though I hung up my toolbelt (in a professional capacity), that thinking never left me.
And now I build digital workflows instead of structures. My materials are different—API tokens, appscripts, and webhooks—but the principles are identical.
BottBott is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
This week, I finished a lead-generation system for my own outreach: contacts flow in from LinkedIn, and Claude analyzes and scores them against a custom rubric, so I can focus on the highest-value leads right away. Next, the generated patterns are matched to likely bottlenecks so I can craft a custom message that resonates with each lead. Lastly, outreach sequences are generated automatically, and Gmail integration tracks LinkedIn connection request acceptances and updates the sheet, so my funnel is continually filling.
I’ll admit, it’s not flashy, but neither is plumbing. But, like good plumbing, the system runs with minimal supervision.
With all the hoopla these days about artificial intelligence killing creativity, I’ve realized something: AI hasn’t replaced my internal craftsman; it’s just given me new tools. The same eye that once spotted a poorly planned material staging area now spots a workflow with too many manual handoffs.
The instinct is identical. Only the medium changed.
Why Systems Thinking Transfers From Job Sites to Operations & Workflow Automations
Back when I was running my construction company, I used to make jigs for everything. Some were simple blocks for repetitive marking or cutting. Others were complex “key”-based jigs for reproducing router lines at regular intervals.
Their purpose was always the same: systematize repetitive tasks so I could focus on craft.
When I moved into digital work—first as a content writer, then into SEO—I was essentially drifting. I started numerous content-writing businesses and passion projects that didn’t amount to much and bounced between strategies. However, one thing remained consistent throughout: the way in which I systematized everything I did. I didn’t realize it at the time, but I was actually building valuable skills: I was learning how information flows, how to structure digital processes, and how to identify bottlenecks.
Every dead-end business venture taught me something about systems. Every frustrating manual process made me think: “There’s got to be a better way to do this.”
Turns out, there is.
A Serendipitous Moment That Changed Everything
My personal breakthrough came through a combination of timing and serendipity. I’ve been using LLMs heavily since early 2023. Not long after I transitioned into content writing, they showed up. And being new and needing an edge, I embraced them wholeheartedly. So, I began using them to streamline real work, building custom prompts for my workflows, testing and iterating on them daily, and learning what breaks and why.
Fast forward nearly three years—and still tunnel-visioned on content writing—I found myself on a call with Kevin Clark, a PCC and executive coach, who suggested I explore doing something with AI. His suggestion came right as I was building an AI agent called Career Pivot Labs, a tool to help me reframe my past work experience into high-level transferable skills.
When I ran myself through the agent intake, it assessed me as a systems thinker. That, plus Kevin’s suggestion, clicked some seemingly disparate pieces into place.
From then on, I stopped seeing my construction background as irrelevant to a remote career. I stopped viewing my floundering content businesses and passion project podcast as a waste of time. Instead, I realized that my past work wasn’t a series of dead ends, but an essential knowledge base and dataset for something bigger.
What Construction Taught Me About Workflow Design
The principles that governed job site efficiency are identical to what make AI workflows effective:
Input → Process → Output → Quality Control
In construction, this looked like:
materials delivery → staged work sequence → finished product → final inspection.
In AI-infused automation workflows, it’s:
data input → structured processing → formatted output → validation check.
The core logic hasn’t changed. I’m still solving the same problems:
How do you identify and prevent bottlenecks?
How do you maintain quality at scale?
How do you build systems that work without constant supervision?
When I built that lead generation system, I wasn’t thinking like a programmer. I was thinking of someone who’s managed complex, multi-step processes where a broken handoff can kill the entire flow.
The gap between “having an idea” and “building an MVP” has shrunk to almost zero, but only if you understand how to structure instructions.
A Step-by-Step Guide to Auditing Your Experience for AI Skills
Here’s the process Career Pivot Labs walks users through, which you can do yourself:
1. Map Your Process Thinking
Look at any job you’ve held and identify:
What were the repeating processes?
How did information or materials flow through your work?
Where did things typically break down?
2. Extract the Logic Patterns
Construction taught me staging and sequencing. Content writing taught me input-processing-output. Customer service teaches handoff management. Retail teaches quality control under volume.
Every job has transferable process logic.
3. Identify Your Automation Instincts
When have you thought “there has to be a better way to do this” while clicking between tabs, copying data, or doing the same task repeatedly? That frustration is valuable; it’s your systems brain identifying optimization opportunities.
4. Connect the Dots to AI Tools
Your process instincts translate directly:
Workflow design → Prompt engineering and agent instructions
Quality control → Output validation and error checking
Handoff management → Tool integration and data passing
Scaling systems → Automation and batch processing
Building Your First Agent (Without Writing Code)
Much to the chagrin of many, you don’t need to be a developer to build digital things anymore. You just need to understand how systems work.
Here’s where to start:
Begin with prompts, not agents. I’ve been using LLMs daily for over three years, completed Google’s AI prompting courses, and built dozens of custom prompts. Start there. Learn how to structure instructions, set constraints, and define output formats. You’ll build intuition for what LLMs can and can’t do reliably.
Use your existing process knowledge. Take something you already do well—project planning, client intake, content review, whatever—and map out every step. Turn that into structured instructions for an LLM.
Test on real work. Don’t build hypothetical tools. Build something you’ll actually use. Theory is useful, but you won’t internalize what works until you’ve debugged it at 11 PM because it’s doing something inexplicably wrong.
Expect iteration. Both prompts and agents will fail in weird ways. The goal isn’t perfection; it’s understanding what breaks and why, so you can decide if the trade-off is worth it.
An Action Checklist So You Can Start Building Tomorrow
□ Audit one work process: Pick something you do repeatedly. Map the steps from start to finish.
□ Identify the bottleneck: Where does this process typically slow down or break?
□ Write it as instructions: Pretend you’re training someone else to do this task. Be specific about inputs, outputs, and quality standards.
□ Test with an LLM: Paste those instructions into your favourite LLM with some real data. See what happens.
□ Refine based on results: Fix what broke. Tighten what drifted. Clarify what got misinterpreted.
□ Build one thing this week: Don’t aim for perfect. Aim for functional.
The most valuable asset in the AI era won’t be a specific technical craft, but the architectural mindset required to direct AI as a digital workforce. As workplace value shifts from the labourer to the systems designer, those who treat GenAI models as the engine and their own intuition as the blueprint will attract the most attention.
And if you’ve recently lost your job—or changed careers as I have—just remember: Your past wasn’t a waste, it’s a valuable knowledge source and future training data.
The question now—beyond whether you can learn to build with AI—is whether you’re willing to recognize that you already know how systems work.
You just need to point that knowledge in a new direction.
The AI era favors the systems thinker—but every thinker needs a builder. If you want to improve your workflows but don’t have time to design prompts and agents yourself, I’d love to help. I build custom AI systems for client-facing and internal operations, applying the same “construction-grade” logic to digital flows. DM me to discuss your workflow, or follow along for more insights on practical AI adoption and the mistakes you only want to make once.
---
If you’re ready to audit your own experience for transferable AI skills, Career Pivot Labs can walk you through the same assessment that helped me see my construction background as an asset, not ancient history. DM me for info.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.