The guy in the blazer had a slide deck with forty-two slides. I know this because the business owner who forwarded me the proposal counted them. Forty-two slides to explain why a fifteen-person landscaping company needed an “AI-enabled customer engagement transformation platform” for somewhere in the neighborhood of $180,000.
The proposal was full of words like synergy and machine learning pipeline and scalable cloud-native architecture. What it didn’t mention was that the company’s QuickBooks data hadn’t been backed up since March. Or that two of their field supervisors were still sharing one Windows login because nobody had set up proper accounts when they hired them three years ago. Or that their email was running on a server under someone’s desk that rebooted itself every Tuesday for reasons nobody had ever investigated.
That server under the desk is load-bearing. The AI chatbot is not.

The Tax on Confusion
Here’s the thing about the current AI moment: it’s real, and it’s also being used to extract money from people who don’t have a strong frame for evaluating it. Those two things are both true at the same time.
The technology is genuinely useful. I use it myself — in automation scripts, in documentation workflows, in ways that save real hours. But useful tools have always attracted a second wave of people who are less interested in the tool and more interested in the confusion surrounding it. The confusion is where the margin is.
“Most of the AI hype aimed at small businesses right now is a tax on confusion. The vendors collecting that tax have, in many cases, never SSHed into a production box in their lives. They’ve read the same breathless articles you have, assembled them into a deck, and added a logo.”
Real modernization is boring first and clever second. It looks like: identity done right, so you’re not sharing passwords and you can actually offboard someone when they leave. Backups you’ve tested by restoring from them — because a backup you’ve never restored is just a feeling, not a backup. Workflows you’ve mapped and cleaned up before you point any automation at them, because automating a broken process just breaks it faster. Vendors who don’t own your data and can’t hold it hostage when you want to leave.
That’s the foundation. You can’t architect your way out of a mess you refuse to look at. And you definitely can’t AI your way out of one.
What Fourteen Years Actually Looks Like
I’m not going to recite a resume at you. But I’ll tell you a few things about how I got here, because context matters when you’re deciding who to trust with your infrastructure.
I started in IT the way most self-taught people do — by being the person in the room who wasn’t afraid to break something to figure out how it worked. Before that, I worked in a Top Secret SCIF monitoring classified military communications systems. Before that, I was catching fish commercially on the Pacific at an age when most kids were worrying about algebra homework. The through-line in all of it: systems have logic, and if something’s broken, you trace it back until you find where the logic failed. Nature doesn’t negotiate. Neither does a broken identity stack at 2am.
For the last several years I’ve been doing senior escalation work across hundreds of managed customer environments at a major enterprise cloud provider. Before that, I led a five-person cloud team through a full hybrid migration covering about fifteen client sites and two thousand end users. I build my own automation in bash, Python, and Rust — not because I’m trying to impress anyone, but because that’s how I think through problems. Writing the script forces me to understand the thing.

I’ve been paged at 2am and fixed it. I’ve also been the guy who told a client they were about to pay $40,000 for a solution to a problem they didn’t actually have. Both of those are part of the job.
The Anti-Pitch
Look, the thing that separates me from the blazer guys isn’t a longer credential list. It’s that I’ll tell you when you don’t need what you’re being sold.
That sounds obvious. It isn’t. Most IT vendors, consultants, and MSPs are structurally incentivized to sell you more — more seats, more licenses, more managed services, more of whatever their preferred vendor is paying them to push this quarter. There’s nothing conspiratorial about it. It’s just how the economics work. The kickback is baked in. The upsell is the business model.
My business model is simpler and less exciting: you pay me for my time and judgment, I give you my honest assessment, and if the honest answer is “you’re fine, save your money,” that’s what you get. I don’t have vendor partnerships. I don’t get referral fees. I’m not going to recommend a $600-a-month tool when a $15-a-month tool and an afternoon of configuration does the same job.
Same rules for everyone. That’s the whole thing. The willingness to say “no” when no is the right answer — that’s not a virtue, that’s just basic integrity. It’s also, apparently, rare enough that it’s worth mentioning.
What I’ll Actually Do
If you bring me in, here’s how it goes. I start with an honest infrastructure audit — not a sales document, an actual picture of what you have and what it would take to make it solid. I’ll look at your identity setup, your backup situation, your vendor contracts, and where your data actually lives and who actually owns it.
From there, I’ll tell you what your AI readiness actually looks like — which mostly means whether your data is organized, accessible, and clean enough for automation to help you. Usually there’s work to do there before anything else. That work is unglamorous and valuable and most vendors skip it entirely because it doesn’t justify a six-figure proposal.
If there are workflows worth automating, we automate them first with the simplest tool that works. Complexity is a liability, not a feature. Then — and only then — if there’s a genuine use case for AI tooling, we talk about that. With honest numbers. With clear ownership. With an exit path that doesn’t require you to beg a vendor for your own data.
I’ll also walk away from engagements where the right call is “stay where you are.” That’s not a sales tactic. It’s the deal.

The Invitation
I’m taking on a small number of contract engagements this year. Small to mid-sized businesses only. Owners who have some skin in the game and are tired of being pitched by people who’ve never had to fix anything under pressure.
If you’re staring at an AI pitch deck right now wondering if it’s bullshit — email me. If your IT situation is somewhere between “held together with hope” and “we think we have a backup somewhere,” email me. If you’ve been told you need a cloud transformation and you just want someone to look at your actual setup and tell you the truth, email me.
admin@scrivener.pro.
I’ll tell you straight. And if the answer is “you don’t need me,” I’ll tell you that too.
— Awasen

Leave a comment