Ilias Gibigaye built a YouTube summarization workflow. He’d been trying to keep up with a growing list of creators — educational channels, business content, podcasts. The manual process was slow, so he automated it.
Five steps: find the video, copy the URL, navigate to Lindy, paste the URL, run the summary. He’d save the output to a notes doc he kept for this kind of thing.
It worked. During the day, he used it regularly.
After 8pm, when he was tired, he’d skip it. Watch the video, let it pass, move on. The workflow that was supposed to compound his learning just sat there while he defaulted to passive watching.
I was coaching him through his Lindy setup when he mentioned this. And the first question I asked was: where do you spend most of your time?
Slack.
The Rebuild
We moved the whole thing. Instead of navigating to Lindy, Ilias now types the URL directly into a Slack channel he’d already created for this kind of content. The trigger fires from there. The summary comes back to the same channel.
No switching apps. No new tabs. No five-step process to remember when he’s already winding down for the night.
The workflow didn’t get more sophisticated. The underlying automation is roughly the same. What changed was where it lives — inside the tool he’s already in, in the channel he already checks.
He hasn’t skipped it since.
Why This Keeps Happening
I’ve seen some version of this pattern repeatedly when working with clients on AI adoption.
Someone builds a workflow they’re genuinely excited about. They use it for a week or two, sometimes a month. Then it quietly goes unused. They assume the automation needs fixing. Maybe the prompt is off, or the tool isn’t right, or the output format needs work.
Sometimes that’s true. But a lot of the time, the actual problem is location.
The output lands somewhere they don’t already check. A new app they have to remember to open. A dedicated folder in their email that fills up unread. A portal built specifically for AI outputs — which means they have to remember the portal exists and then remember to go there.
This is classic friction. And friction kills habits even when the underlying behavior is something people actually want to do.
Think about the last habit you tried to form that didn’t stick. Was it really that you didn’t want to do it? Or was there an extra step — something you had to set up first, somewhere you had to go — that made it easy to skip when you were tired or distracted?
Workflows are habits. They follow the same rules.
Work Where You Already Live
The principle I use now when designing any AI workflow: the output should land where the user already spends time.
If they live in email, the summary should arrive in their inbox. Not in a separate notes app that syncs to email. In their actual email inbox, in a folder they already check.
If they live in Slack, the output should land in Slack. A channel they already have open. Not a new channel they have to think to check — in Slack itself.
If they primarily work on their phone, results should come as a text or a push notification — not a web app they have to navigate to.
The corollary: don’t design workflows that require new habits just to access the output. If someone has to form a new behavior just to see the results of your automation, you’ve added a second adoption problem on top of the first one.
Where I’ve Seen This Go Wrong
I worked with a team that built a genuinely useful meeting summary system. After every call, the agent would generate a structured recap — action items, decisions, key context — and post it to a shared Notion workspace.
In theory, great. In practice: the team still worked primarily out of Slack. Nobody checked Notion consistently. Meeting summaries accumulated, unused.
The fix wasn’t to rebuild the summarization. It was to route the summaries to a Slack channel the team already used for follow-up tasks. Same content, different destination. Adoption jumped immediately.
A Practical Way to Audit Your Current Setup
If you have AI workflows that aren’t getting used consistently, run through this quickly:
Where does the output of each workflow land? Is that somewhere you check every day without thinking?
If the answer is no — if the output goes to a destination that requires intentional effort to access — that’s probably where the adoption problem lives.
The fix is usually simple: reroute the output to somewhere you already are. Your main email inbox. Your most active Slack channel. Your phone’s messaging app.
This is the whole “One Tweak a Week” idea in practice. You don’t need to rebuild anything. You just need to move where the output appears.
That one change — where it lands — is often the difference between an automation you use daily and one you check in on occasionally and feel vaguely guilty about.
Thanh Pham is the founder of Asian Efficiency. He coaches individuals and teams on building AI workflows that actually get used.
Recommended for you
Want the full system? 25X is the flagship productivity system we teach.
Explore 25X →