The Unpopular Opinion
Why Your First AI Project Should Be Boring.
I've built the shiny tools. Some of them worked beautifully and still didn't matter. The projects that pay are the unglamorous ones — and I have the scar tissue to prove it.
Quick Answer
Why should your first AI project be boring?
Because boring projects solve problems that already cost you money every week, which means they pay for themselves and actually get finished. A shiny project is built because it's exciting; a boring project is built because the problem is real. A useful system quietly creates value every day — a shiny system nobody needs becomes another expense. Pick the repeated task that slows everyone down, let AI bring it 90 percent of the way, and keep a human on the final 10 percent.
- →Shiny projects are built for the builder; boring projects are built for the business
- →A useful system quietly creates value every day
- →A useful feature is not always a complete business — the voice-notes lesson
- →The 90/10 rule: AI brings the work 90 percent, humans keep the final 10
- →Find boring problems where work is repeated, forgotten, or done differently every time
The Confession
I've built the shiny tools. Here's what it taught me.
I have built many shiny tools that looked good and did exactly what I intended them to do. The problem was that I built them because I wanted to build them, not because customers were asking for them.
That's a hard thing to admit as someone who builds technology for a living. The engineering was fine. The demos were impressive. But I was answering questions nobody had asked — and a perfect answer to a question nobody asked is worth nothing.
I see the same pattern in owners adopting AI. The exciting project — the chatbot, the content engine, the AI receptionist with the polished voice — gets the budget. The boring project — the quote follow-up, the intake form, the document processing — keeps leaking money in the background. One of those is a hobby. The other is an investment.
A useful system quietly creates value every day. A shiny system that nobody needs becomes another expense.
— Steffen deGraaf, BotLogix
The Counterpoint
Boring on the backend can still be excellent for the customer.
Boring doesn't mean the customer notices nothing. In 2018, we built an early keyword-triggered DM automation. In our experience with that campaign, it brought in approximately 500 new contacts within 24 hours.
The backend was deeply unglamorous: keyword matching, routing rules, message queues. Nothing you'd put in a keynote. But the customer experience was excellent — someone commented a single word and got an instant, relevant, personal-feeling response. What made it satisfying wasn't the number. It was watching a complicated backend system turn into an effortless experience for the customer. The technology almost disappeared.
That's the right definition of boring: unglamorous plumbing, invisible to the customer, felt only as speed and reliability. The best AI systems are boring where you build them and excellent where the customer touches them.
The Hard Lesson
The voice-notes product that worked perfectly — and still failed.
I built a tool that turned voice notes into structured action items. Record a thought between appointments, get back organized tasks in the right workflow. It worked. It worked exactly as designed, every time.
It failed as a standalone product. Not because the technology was bad — because a useful feature is not always a complete business. People liked it. They didn't need it badly enough, often enough, to pay for it on its own. The same capability, embedded inside a workflow a business already depends on, is genuinely valuable. Standing alone, it was a demo.
The lesson transfers directly to your business: judge an AI project by the problem's frequency and cost, not by how well the demo goes. A capability that removes real friction from a daily process will get used. A capability that merely impresses people will get a compliment and a closed tab.
The Safety Rule
The 90/10 rule: AI prepares. We approve.
Boring projects have another advantage: they're safe to automate most of the way. The pattern I build to is simple. AI can often bring the work 90 percent of the way there. The final 10 percent should remain human.
AI drafts the quote response; you approve it. AI extracts the contract details; you spot-check the flagged fields. AI triages the inbox; you decide what the angry customer gets told. That final 10 percent protects the company's reputation, brand, accuracy, judgment, and voice — and it's exactly what keeps a boring project boring, in the best sense: predictable, controlled, no surprises reaching a customer.
Four words worth writing on the wall of every project: AI prepares. We approve.When a proposal skips the approval step to save ten minutes, that's not efficiency — that's risk wearing an efficiency costume.
AI can often bring the work 90 percent of the way there. The final 10 percent should remain human.
— Steffen deGraaf, BotLogix
The Hunt
Where to find your boring problem.
You don't need a workshop to find the right first project. You need to notice three kinds of tasks — they're already in your building.
The repeated task that slows everyone down
The one that comes up in every staff complaint. Re-entering the same details into three systems. Chasing the same approval every week. If your team jokes about it, it's a candidate.
The task that gets forgotten
Follow-ups that slip. Quotes that sit. Check-ins that only happen when someone remembers. These don't feel expensive because the loss is invisible — but forgotten follow-up is where revenue quietly disappears.
The task done differently every time
Ask three people how intake works and get three answers. Inconsistent process means inconsistent output — and it's impossible to improve, because there's no single version to fix. Standardize it first, then automate the standard.
Pick one. Map it. Let AI take the middle 90 percent. Finish it, measure it, and put it into operation. Then — and only then — go find the next boring problem. The companies winning with AI aren't the ones with the most impressive demos. They're the ones with the most finished, unglamorous, quietly profitable systems.
Go deeper
Using AI in Your Business
The owner's guide for 5–25 person companies.
What to Automate First
The framework for picking your first project.
The 7 AI Mistakes
What sinks small business AI projects.
AI Implementation Costs
Real numbers, not sales-deck ranges.
5 AI Projects Every SMB Should Run
Specific projects with real ROI.
Workflow Proof Sprint
Prove one boring workflow before you build big.
30-minute strategy session
For real businesses with real problems.
Not for tire kickers. Not for tech tourists looking for a demo. For business owners who already know something in their operation is broken — and want to know if AI can fix it.
This isn't about replacing people.It's about killing the bad business processes your team hates as much as you do — the manual work, the missed calls, the after-hours admin.
Strategy sessions are free. Strategy Days are $1,000, capped at 4 people per session. Larger teams or multi-location workshops by special arrangement — ask when you book.
Your session is with
Steffen deGraaf
Founder, BotLogix · Burlington, ON