“Who approved that post?”
It went live with a wrong number in the headline. Nobody remembers signing it off.
content: A publishing pipeline that never fakes a green.
ConvOps runs your content pipeline as governed AI agent workflows. Briefs become drafts. Drafts face two reviews at once. Only a double green ships. Put it on a schedule and it runs every morning, remembering what yesterday taught it.
claude cowork · two reviews side by side · runs on a schedule
new chat
task
step 1 of 6Post: onboarding checklists
brief recall draft two reviews approve approvalpublish
the problem
Fast content. Nobody knows what shipped.
AI drafts in seconds. Then the voice, the facts and the review depend on who was in the chat.
“Where did that decision go?”
The team agreed on no jargon in March. Every new chat writes jargon again.
“What happens when someone leaves?”
Your editor knew the voice. Now the AI and the new hire are guessing.
watch it work
One article, brief to published.
A real post, run through the Blog Post workflow from Claude Cowork. Then put on a schedule.
01you, in Cowork
Ask for the post.
You ask for next week's post in plain words. ConvOps creates the task and attaches the Blog Post workflow, with its reviews and its approval.
example
Write next week's post on onboarding checklists. Readers are ops leads.
task created · Blog Post workflow attached
Brief captured. Recalling our voice rules before I draft.
02the AI
It recalls the voice.
Before drafting, the AI asks the brain: the reader, the voice rules, the rulings from past reviews. Run 40 writes with what run 1 had to be told.
example
askblog post voice and audience
memoryReaders are ops leads, not developers. No jargon.
memoryHeadlines under 60 characters. Ruled in the March review.
routeBrand voice guide: docs/brand/voice.md
03the AI · research and draft
Sources first, then words.
Sources are gathered and pinned before the draft starts. Every claim in the draft points back to one of them.
example
Onboarding checklists people actually finish
source 1Why most checklists stall after day one
source 2One owner per step, named up front
source 3Steps small enough to finish in one sitting
3 sources pinnedvoice rules applied
04two reviews
Two reviews. One red.
SEO and engagement review the draft side by side. A finding sends it back to the draft until both are green. A hard red files a bug instead of shipping.
example
- Keyword in the title
- Headline under 60 characters
- Internal links
- Answers the reader early
- Concrete examples
- One clear next step
05you approve
You release it.
Keep the approval gate and the pipeline drafts on its own but ships only when you say so. Remove it when the double green is enough.
example
waiting for you
Publish "Onboarding checklists people finish"?
SEO green · engagement green · 3 sources
Send backPublish
approvedPublish step unlockedThe sources travel with the post.
06the schedule
Tomorrow it runs itself.
Put the workflow on a schedule. Each morning a new task fires with the whole process attached. You stop starting it and keep governing it at the gate.
example
every weekday · 07:00next fire · tomorrow 07:00fired
- Post: first-week goalsdrafting
- Post: handover noteswaiting for you
- Post: onboarding checklistspublished
example
Write next week's post on onboarding checklists. Readers are ops leads.
task created · Blog Post workflow attached
Brief captured. Recalling our voice rules before I draft.
the workflow
The Blog Post workflow, running.
From brief to published in six steps. Reviews run side by side, and you keep the last word.
workflow · Blog Postrunning · step 1 of 6
- 01
Brief
Topic, reader and angle, captured as structure, not a doc.
- 02
Recall
Voice rules and past rulings come back before a word is written.
- 03
Research and draft
Sources gathered first. The draft cites them.
- 04
Two reviews at once
SEO and engagement side by side. Findings send it back.
decision - 05
You approve
Keep this gate and nothing publishes without you.
needs a person - 06
Publish
Ships with its sources attached.
what it remembers
The voice lives in the brain.
Every review makes the next draft better. Kept by the team, not by one editor.
Voice rules
Readers are ops leads. No jargon. Plain verbs. Recalled before every draft.
how we sound
Past rulings
Headlines under 60 characters. Decided once in review, applied to every draft after.
decided once
Where sources live
The research folder and the brand guide, so no draft starts with a search.
where it lives
Linked to each post
Every ruling points back to the post that produced it.
traceable
the difference
Prompting vs a pipeline.
- A new chat for every post, the voice explained again
- One reviewer, when someone has time
- A wrong claim ships because nothing stopped it
- Someone has to remember to start it
Reviews side by side
The review step tells your AI session to start both reviewers at once. ConvOps holds the step; your session runs the two agents on the same draft.
The loop is a detour
Findings send the run into a fix sub-workflow that returns to the draft step. It repeats until the reviews pass.
Two schedule modes
An RRULE schedule either creates a fresh task on each fire, or resumes the same task where its workflow stopped.
Rules reach the step
Policies like "every claim carries a source" arrive with the step the agent is on. Guidance at the moment it acts.
start this week
Four moves to your first run.
Connect the AI tool your team already uses, install the workflows, and run one real piece of work through them.
- move 1
Install the Content Studio pack: blog post, social post, newsletter, repurpose
- move 2
Run one article attended: brief, research, draft, two reviews in parallel
- move 3
Watch the fix loop send findings back to the draft until both reviews pass
- move 4
Then add the schedule, and stop starting it yourself
install these
marketing
Blog PostTake a blog post from brief to published draft. The AI drives through Outline, Draft, Edit, and a final read-through.marketing
Social PostTake a social post from brief to ready-to-publish. The AI drives through Draft Variants and a Polish checkpoint.marketing
Newsletter IssueBuild one newsletter issue from theme to send. The AI drives through Curate, Draft, and a send-review checkpoint.marketing
Content RepurposeTurn one piece of content into several formats. The AI drives through Angles, Produce, and a review checkpoint.built on
The parts doing the work.
AutonomySchedules that start the work for you.The brainVoice and rulings, recalled every run.Workflow engineReviews, loops and gates.TemplatesBlog post, social, newsletter, repurpose.
ConvOps is the operations layer for AI agents: an MCP server that holds your team's process as workflows, with approval gates, a shared memory and an audit trail. It runs no AI models.
questions
What teams ask.
Is this actually autonomous, or a fancy scheduler?
The schedule only starts it. The loop then recalls what past runs learned, drafts, faces its reviews, and stops at the gates you placed. Selection, memory, verification, and governance are what make it a system rather than a timer.
What stops it from publishing something bad?
The verification gate. Reviews run in parallel and findings loop the draft back until it passes; on a hard red the workflow files a bug task instead of shipping. A loop that cannot fake green is the entire point.
Where does it keep our style and standards?
Two places, deliberately: policies deliver rules like "every claim carries a source" into the active step, and memories carry the accumulated rulings, so run 40 writes with everything run 1 had to be told.
Can a human stay in the loop?
Wherever you want one. Keep the approval gate before publish and the pipeline drafts autonomously but ships only when you release it; remove it and the double-green reviews are the bar.