Let's fix this process... by not adding AI.
How we turned a print-and-memory ritual into a calm, trackable workflow before touching AI.
They told me they wanted automation. What they had was a ritual.
Emails arrived all day. Critical messages mixed with noise. The team kept things moving by printing, highlighting, stacking papers, and relying on one person’s memory to keep priorities straight.
It worked for a long time.
Then the volume crept up, urgency increased, and the cracks started to show.
The owner had built the company over decades. The team knew the aviation industry inside out. They were fast, practical, and cautious about new technology for good reason.
But one of the key persons was spending 2 hours a day triaging emails. They needed breathing room.
I offered them relief.
Processing power? Not the problem.
The real problem was invisible work. In their world, work only became “real” when someone printed it, left it unread as a reminder, or added a post-it to some dark corner of their brain for later.
Think a very physical kanban board, but instead of cards, printed sheets of paper, and instead of columns and swimlanes, people’s desks. In different physical locations.
That is not a reliable workflow, especially when there are remote employees involved.
The moment you try to automate a setup like that, everything feels fragile. Where did that paper go? How did you decide it needed to go there? How do you know the person actually read it?
So my job became clear:
Make work visible, owned, and traceable.
Then remove manual handling without breaking the pace that keeps their clients happy.
The mindset shift
If you want to help a team like this, you do not begin by telling them to work differently. They’ve been successful and profitable for decades. If you parachute in and say “you should do X”, they will tell you to buzz off - and rightfully so.
If you’re serious about implementing change, you must begin by changing what they can see.
What’s the smallest possible change with the biggest impact?
In this case, it was making all pieces of paper, from all desks, visible. Everyone knows where each paper is, who’s working on it, and what happens next.
Your basic task management system, fed from the mailbox that fuels the business.
We did it with tools they already used, inside the environment they already trusted. No Trello, Jira, ClickUp. The last thing they need in this scenario is to add the psychological pressure of learning a new tool.
So the first constraint was: No new platform, no extra logins, no culture war.
Then we removed the worst manual work
Most teams mess up when they try to automate judgment. We automated mechanics only. All the repetitive handling that eats hours and adds risk of human error:
Saving attachments
Copying documents into the right folders.
Forwarding the same patterns of emails to the same people.
Printing as a way to remember.
Running the same “check this, then route that” routine dozens of times a day.
But the key for success was a simple decision we made early on:
Humans stayed in charge where judgment mattered.
This gave them time to stop chasing the work. They audited the work instead.
They watched how items moved. They corrected the system when it was wrong. They confirmed it was safe to trust.
That one shift changed the tone in the room because change and adoption stopped feeling like a fight. It turned into a recognition of their knowledge and expertise, and delivered quality control.
Paper stayed. The chaos did not.
They still needed paper. In aviation, compliance still matters, audits still happen. People still have to hold a physical file in their hands.
So we did not try to remove paper, we changed its role. Instead of using printed paper as work items, when work is finished, the system produces a complete packet for filing and compliance.
Paper stayed, but it stopped driving the day. And that gave them a bunch of perks: the team moves faster, with less stress and fewer errors. Their main operator who spent 2 hours a day triaging the inbox can now take holidays without the whole business stalling, or the owner having to jump in.
Why I decided to write about this: The part I am proud of
If you’ve been reading my articles for a while, you’ll know that I’m becoming allergic to large tool stacks.
So my stance on this project was: build a system without adding overhead. It was a real systems project, as Donella Meadows would say:
"If you're interested in changing that system, you will likely have an interest in changing a feedback loop of one sort or another, either decoupling a feedback loop to allow a system to move in a different direction or perhaps reinforcing a feedback loop which might push a system in another direction". - Donnela Meadows
The technology was straightforward.
The real work was building something that fits how the team already thinks, respects their experience, does not slow them down, does not put the business at risk on day one, AND relieves key actors of boring, repetitive work.
The results we gathered: what actually changed
The owner no longer had to be the default bottleneck.
The main operator saved 2 hours a day.
Work stopped hiding in paper piles and unread messages. Everything was visible to everyone.
But for me, the most beautiful part is, the company gained something it had never fully had before: A real operational spine, ready for the future.
Now when they bring up AI, it can be leveraged for real. Instead of going for it as a call for help, it will speed them up further and free their time for more real and relevant business.
The lesson
I must repeat this: Automation is not about stacking tools.
It is about turning “how we do it in our heads” into something a system can execute every day and you can audit the day after.
You start with a clear system that is intentional and makes work visible and controlled.
Then you automate the boring mechanics.
Then, and only then, you let AI reduce the remaining cognitive load.
If you are considering implementing AI, ask yourself one thing first:
Can your team already see every piece of work, know who owns it, and trust the system more than their memory?
If the answer is no, that is the real project.
AI comes after that.
Happy building,
-R.