I make videos. The part with the timeline and the effects is fairly easy to show someone. The less glamorous part is figuring out what happens before the edit starts—and remembering what still needs to happen after it ends.
At American Public Life (APL), a request might show up in Teams, move through a review link, and eventually land on Wistia. Those steps made sense while I was doing them. They were much harder to see from the outside.
I usually knew where a project stood because I was involved in it. But answering a simple question could mean piecing together its history from several places. Not my favorite kind of detective work.
I started designing a video workflow in Monday.com, with AI helping me sort through the process. A beautifully organized board would be nice, but only if it could survive contact with an actual video project.
Making the process visible
Monday.com was already available, but we hadn’t had time to explore it thoroughly or build something specifically for video. I had to figure out what the platform could do without letting its features decide how we should work.
I organized the design around three connected areas:
Request Intake → Video Production → Video Library
A video should have a path from the original request to the finished asset. And “finished” should mean we can still find the correct version when someone asks for it later.
Project Snapshot
Company: American Public Life (APL)
Project: AI-assisted process mapping and video workflow development in Monday.com
My role: Production subject-matter expert and workflow designer
Production environment: Monday.com, internal AI tools, Teams, Outlook, Wistia, Adobe Creative Cloud, online review tools, and shared storage
Project status: In implementation and refinement
The Context
An approved script is a wonderful starting point. But plenty of projects begin with something closer to, “Hey, can we make a video about this?” Sure! Let’s start with what “this” is.
The kind of video changes the process. A product-education piece that needs compliance approval follows a different route from a quick internal clip. And even when the edit is approved, there may still be work to do before anyone can use it.
I’d gotten used to carrying those differences around in my head. It worked well enough for me. It was less helpful when someone else needed to see the whole process.
The Challenge
I could have started with a standard production template. Instant organization! Except that a template which makes every video follow the same path can create more work than it saves. I needed the structure to fit the projects, not the other way around.
The first tricky decision was what belonged in the project record. I wasn’t trying to move every conversation or production file into Monday.com. I wanted enough information in one place that someone could tell what was happening without retracing the entire project.
Approvals made that especially important. Sometimes compliance needs to review a video before it can be published. Sometimes a much smaller review is enough. The board had to make room for both without treating an internal clip like a major production.
And I wasn’t about to ask everyone to abandon email or Teams. No one should need a crash course in Monday.com just to comment on a script. That isn’t an improvement. It’s homework.
How I Used AI
I gave AI an unusual assignment: interview me about my own job.
I explained how our video projects worked and let it ask follow-up questions. What happens when a request comes in without a clear scope? Why does one video need a compliance review while another doesn’t? What happens to the final version after we publish it?
I knew many of the answers without thinking. Then a question would come along that made me realize I’d been handling exceptions on instinct. I could make those calls in the middle of a project, but explaining exactly why I made them took more thought.
AI helped turn those conversations into possible ways of organizing the work in Monday.com. I checked its suggestions against actual projects before deciding what to use. A few were impressively complicated solutions to problems I wasn’t having, so I left them out.
That back-and-forth let me see assumptions I hadn’t noticed I was making. The knowledge came from doing the work. AI helped me question it.
My Role
Because I’m involved in video projects beyond the edit itself, I could test the proposed workflow against what actually happens during production. I mapped the process and worked out how it could live in Monday.com without adding steps just for the sake of a tidy board.
I tried the structure against different kinds of video projects and worked with colleagues to review it. I kept coming back to one question: could someone understand where a project stood without needing me to explain the entire backstory?
What I Designed
Request Intake
The intake section starts with what we know. A request might be a complete brief, but often it’s a conversation. I designed a place to capture enough of that conversation that the missing details won’t be a surprise halfway through production.
Not every question has to be answered right away. I’d just rather know what’s missing at the beginning than find out when I’m already deep into an edit.
Video Production Board
I used a Scrum-inspired structure so a request could sit in a backlog before it became active work. That sounds simple, but there’s a real difference between “someone asked for this” and “we’re making this now.”
Once a project is underway, the stages can flex around what it needs. A short internal video shouldn’t be stuck waiting in a compliance queue that doesn’t apply. A product video that does need approval should have that step built in.
The board is designed to show who owns the next action and what’s holding things up. Ideally, “Where are we on this video?” won’t require a tour of everyone’s inbox.
A Better Definition of “Done”
The edit is approved. Great! We’re done!
Well, except for the captions. And the Wistia upload. Compliance may still need a final check, and someone has to make sure the right version gets published.
So, maybe not quite done.
I defined completion to include the work that happens after creative approval. If a video needs a form-number check or a future review date, that needs to be accounted for too.
Those last details are easy to remember while they’re right in front of me. The checklist is there for the moment when they aren’t.
Video Library
“Didn’t we already make a video about that?” is a familiar question. The trickier follow-up is, “Can we still use it?”
Finding a video file doesn’t tell me whether its content is still current. The library is designed to keep that information with the final link, so a video needing review doesn’t look just as ready to publish as one that’s cleared for use.
It also gives us a place to find the correct publishing location later. A final file isn’t much help when nobody knows which version actually went live.
Views and Possible Automations
I may need to see all the production details. A manager probably just wants to know whether the project is on track. Different views can make the same record useful to both of us without creating a second round of status reporting.
There are routine handoffs that might make sense to automate, such as letting a reviewer know a video is ready. Those are ideas to test as the workflow develops, not features I’m claiming are already running.
I could happily disappear down an automation rabbit hole. I’m trying to find out what actually needs automating before I move in.
Outcome and Impact
The workflow is still being implemented and refined, so I don’t have adoption numbers or time-savings metrics to report.
What I do have is a clearer map of the work. I can see where information tends to get separated from a project and where a handoff needs more thought. There’s now a structure we can test and improve instead of starting from a blank board.
Personally, I like having the process somewhere other than my head. I can put it in front of someone, walk through it, and notice the awkward parts instead of just working around them out of habit.
Using AI this way was unexpectedly useful. It kept asking what happened next, including when I thought I’d already explained it. Sometimes I had to stop and think. Which, as it turns out, was the whole point.
