From Approved to Published: How Zo.E Handles Distribution
2 min read · 461 words ·
Most people who automate content generation still post manually. They'll have AI write the copy, maybe even format it — then open LinkedIn, paste it in, and hit publish themselves. That's not automation. That's AI-assisted typing.
The whole point of building a system is that the system closes the loop. Zo.E does.
What happens after you tap Approve
WF4 is where James touches content. He gets a Telegram message — the post preview, platform, scheduled time. He taps Approve. That's it. He's done.
What happens next runs without him.
The approved record sits in NocoDB with a flag set. WF5 — the Content Scheduler — runs on a schedule and checks for approved posts ready to go. When it finds one, it hands it to Upload Post.
Upload Post is the layer that actually speaks to LinkedIn and Bluesky. It handles the platform-specific formatting, the API calls, the character counts, all of it. Zo.E never touches those APIs directly — that's Upload Post's job. The split keeps things clean. Zo.E does the thinking; Upload Post does the dispatching.
The post goes live. James doesn't see it happen.
When you want to publish right now
The scheduled path is for planned content. But sometimes you've got something time-sensitive — a response to a news story, a quick observation you want out in the next ten minutes.
WF8 (Adhoc Publishing) and WF10 (Adhoc Content) handle that via webhook. Same end result: Upload Post distributes to the platforms, no manual posting required.
Two paths, one destination. Neither of them involves you logging in.
Telegram as the only interface
This is the bit people underestimate.
When your only interface to the publishing system is a Telegram approval tap, you can't drift into platform use. You're not "just checking notifications" and ending up in the feed for twenty minutes. You approve or reject — and you put your phone down.
WF6 pulls engagement reports on a schedule. That's when James sees numbers. Not by logging in, not by scrolling — by reading a structured digest. The platforms become data sources, not environments you inhabit.
That detachment is a feature, not a side effect. It's actually what you're building toward.
The simplicity at the end
There's something slightly funny about how straightforward the publishing piece is, given how much happens earlier in the system — research, drafting, critique, refinement, human review.
And then at the end: NocoDB flag checked, Upload Post called, post published. Done.
That's the right shape for automation. The complexity lives where judgement is needed. The mechanical part — actually getting the content onto the platform — should be invisible. With Zo.E, it is.
Ta,
James
Founder | Nunlimited
