I’ve all the time been the person who would somewhat spend 4 hours writing a script than one hour fixing the precise downside. Not as a result of the script is quicker, however as a result of you already know the identical difficulty goes to return again subsequent week. And the week after that.
That mindset just about outlined how I ended up the place I’m immediately. From IT admin, to writing full-time, to now constructing DaemonLayer, an automation platform for MSPs.
On this article, I wish to share a little bit of the story behind that journey.
I’d Somewhat Script It Than Repair It Twice
I began my profession as an IT specialist, supporting round 150 customers. From the start, my strategy was easy, if I needed to do one thing greater than as soon as, I wrote a script for it. Password resets, consumer onboarding, monitoring, file server cleanup, if it may very well be automated, it was automated.
Not as a result of we have been drowning in work. Two individuals for 150 customers is completely doable. However automating the repetitive stuff meant we had time to work on issues that really mattered. And for me, that shortly became constructing software program.
It began with scripting, however in some unspecified time in the future the corporate wanted an internet software that the seller wished manner an excessive amount of cash for. So I constructed it in-house. That was over 15 years in the past. That software grew into one of many core merchandise of the corporate, ultimately with two full-time builders engaged on it.
That’s actually the sample that caught with me. Automate the boring work, liberate time, and use that point to construct one thing higher. By the tip, I used to be working your complete IT operation by myself, spending perhaps 20 hours per week on it. The remaining went into improvement and extra proactive work.
From Scripts to Articles
Round that very same time I began writing on LazyAdmin.nl. Not as a result of I had an excessive amount of free time, however as a result of I stored working into the identical factor, I might spend hours determining a repair or constructing a script, and I knew different admins have been coping with the very same downside.
So I began sharing what I realized. PowerShell scripts, Microsoft 365 configurations, networking guides, sensible stuff for IT professionals that simply wanted to get issues working.
That was over 10 years in the past. The weblog grew from a few articles to over 500. At its peak, LazyAdmin had over 700,000 month-to-month guests. That was sufficient to give up my IT job and begin writing full-time.
For the primary time I used to be capable of make a dwelling doing what I actually loved, sharing information, serving to individuals resolve issues. The writing and the scripting all the time went hand in hand. I figured one thing out, made it reusable for myself, after which wrote about it so others may use it too.
Then AI Modified All the pieces
About two years after going full-time, AI grew to become mainstream. And particularly, AI Overviews in search outcomes began taking up.
You’ve in all probability seen them your self. You seek for one thing, Google provides you the reply proper there on the high. Summarized from articles like mine. No must click on via anymore.
Visitors dropped. Not only for me, I hear the identical from just about each fellow author. LazyAdmin went from 700,000 to round 150,000 month-to-month guests. The development is obvious and it’s not going again.
Word
I’m not saying AI is dangerous. I take advantage of AI instruments each single day. However the present mannequin the place search engines like google summarize your content material as an alternative of sending individuals to it’s a actual downside. The individuals who create the content material that AI learns from can now not make a dwelling doing it. I didn’t begin writing for the cash, I began as a result of I wished to assist individuals. However with the ability to pay the payments with it’s what allowed me to do it full-time. That half is breaking.
I nonetheless write. I’ll maintain writing. However lower than earlier than, and with a unique focus.
Why I’m Constructing DaemonLayer
Once I look again at my profession, it’s just about the identical sample each time. See repetitive work, automate it, transfer on. First with scripts for my very own IT setting, then by turning these fixes into articles for different admins.
Now I’m taking that very same strategy and constructing it right into a product.
DaemonLayer is an automation platform for Managed Service Suppliers. It processes help emails, triages tickets utilizing AI, and runs automated workflows to deal with frequent IT requests, password resets, consumer onboarding and offboarding, account administration, and not using a technician having to do it manually.
The issue it solves is one I do know very well. MSPs obtain a whole lot of help emails every single day throughout dozens of shopper organizations. Technicians spend an enormous chunk of their time simply studying emails, determining which shopper it’s for, after which performing the identical repetitive duties again and again.
That’s precisely the sort of work I used to be automating 10 years in the past. The distinction is the dimensions.
The Actual Problem, 100+ Consumer Environments
Once I was automating my very own IT setting, I knew each element. The naming conventions, the group construction, which mailboxes have been shared. I may write a script that labored completely as a result of it solely needed to work in a single place.
MSPs don’t have that luxurious. They handle 50, 100, typically 200+ shopper tenants. Every one is configured otherwise. Totally different naming conventions, completely different license setups, completely different safety insurance policies, completely different instruments.
You’ll be able to’t simply write a PowerShell script and run it in all places. The system wants to grasp every shopper’s setting and adapt to it. That’s the true engineering problem, and actually, it’s what makes constructing this actually fascinating.
Managed Automation, Not Autonomous Brokers
There’s numerous hype proper now round AI brokers. The concept is that you just give an agent a job, hand it bunch of instruments, and let it determine all the pieces out by itself. It makes for nice demos.
However we’re speaking about manufacturing environments right here. Actual shopper tenants. Actual consumer accounts.
If in case you have ever written a PowerShell script that chosen too many customers due to a nasty filter, you understand how quick issues can go unsuitable. And that’s a script you wrote your self, doing precisely what you advised it to. Now think about an AI agent making an attempt each device it may possibly consider to get a job finished, in a shopper’s manufacturing setting.
That’s why I haven taken a unique strategy with DaemonLayer. Each workflow does one particular, well-defined job. The AI handles the triage half, studying the e-mail, understanding what’s being requested, figuring out the suitable shopper and consumer. However the execution follows a managed path with clear guardrails.
Earlier than a password will get reset, earlier than a consumer will get onboarded or offboarded, the system can pause and look ahead to approval. Human-in-the-loop shouldn’t be a limitation, it’s the way you construct belief. Particularly when you find yourself working in another person’s setting.
It’s the identical precept as working -WhatIf in PowerShell. You validate the end result earlier than you execute. The one distinction is that as an alternative of a dry run on a script, you might be reviewing an AI-triaged workflow earlier than it touches a shopper’s Microsoft 365 tenant.
Wrapping Up
The sample has all the time been the identical for me, automate the repetitive work so you possibly can concentrate on what really issues. The medium modified through the years, from scripts to articles to now a platform, however the concept behind it didn’t.
I nonetheless write, I nonetheless share what I be taught. Simply in a unique type now. If you’re enthusiastic about what we’re constructing, check out DaemonLayer.
If in case you have any questions or simply wish to share your personal story, drop a remark under.
