Case study · our own production host

A content pipeline where nothing publishes without passing QA

Four agents turn an approved topic into a published post. The one that verifies the work does not know what the one that wrote it intended — and the one that ships refuses to post anything unverified.

Get my free blueprint
4
agents in the chain
457
drafting sessions
2
publish cadences per week
6
documented runbooks

Read directly off the host on . No estimates.

Why "AI writes your blog" usually fails

The standard setup is one agent with a prompt that says "write a post about X" and a human who reads it before publishing. It fails in two predictable ways. Either the human becomes the bottleneck and the cadence dies, or the human stops reading carefully and the quality drifts until something embarrassing ships.

The problem is that drafting and verifying are the same actor. An agent that writes a post and then checks its own post will pass it, because it already believes the post is correct. You cannot prompt your way out of that — it is a structural conflict of interest.

The chain of custody

We split the work across four agents, each with its own mandate and its own hard limits:

AgentOwnsExplicitly does not
KnowledgeWhat is working per platform, what changed, who to watchWrite anything. It curates.
StudioProducing the draft from an approved topicDecide what to write — that is approved upstream
QAVerifying the artifact against the manifestKnow what Studio intended the result to be
PublisherGetting verified work onto the platformsDecide what publishes, or verify content itself

The drafting agent's instructions are blunt about the boundary: "You don't decide WHAT to write — operator approved that (or a scheduled blog cron picked a non-duplicative topic from approved niches)." Note the second half — even the autonomous path is constrained to pre-approved niches and checked for duplication, so unattended runs cannot wander.

The publisher's three refusals

The shipping agent is the last line before something becomes public, so its limits are written as refusals rather than preferences. Verbatim from its own instructions:

That third one is the least obvious and the most valuable. A publisher that quietly succeeds on two platforms and fails on the third looks green on every dashboard while a third of your distribution silently stops.

The runbooks

Each stage is a documented procedure the agent follows, not a freeform instruction. The drafting agent alone carries six: seed, draft,editor-review, humanize, post,publish-blog. Verification carries one — verify — and it is the same procedure every time, which is the point.

Writing the procedure down and having the agent execute it is what makes the output consistent across hundreds of runs. It is also what makes a failure debuggable: you know which step ran and what it was supposed to do.

What runs unattended

JobCadencePurpose
Primary blog publishTue + Fri, 09:00Topic → draft → verify → ship
Second content trackMon + Thu, 10:00A separate publishing stream on its own schedule
QA daily sweepDaily 11:00Verification pass over pending artifacts
Publisher daily checkDaily 10:00Confirms the shipping path is healthy
QA-to-publisher watcherContinuousMoves verified work forward automatically
Platform knowledge refreshWeekly Sat 06:00Updates the per-platform playbook

The drafting agent has logged 457 sessions, the most of any agent in our fleet. That is the volume the separation of duties has to survive — and the reason we would not run this without an independent verification step.

What broke — the honest part

Both publishing jobs were in a failed state at survey time, and the causes are worth showing:

The lesson we take into client builds from this: wire the last mile before you celebrate a green run. A job that completes successfully and delivers nowhere is invisible in every model-level metric and completely useless to the business.

What this case study does not prove

What we would carry into your build

Three transferable decisions, none of which require four agents: keep the agent that verifies ignorant of what the agent that drafted intended; write the shipping agent's limits as refusals with no override; and treat "delivered nowhere" as a failure, not a success.

If you want a publishing cadence you do not have to babysit, that is the shape of it.

Questions

Can AI agents publish blog content without a human reviewing every post?

Yes, if drafting and verification are separated. A single agent that writes and then checks its own work will always pass it — that is a structural conflict, not a prompt problem. In our pipeline the verifying agent reads the manifest itself and does not know what the drafting agent intended, and the publishing agent refuses to post anything unverified. A human still approves the topic.

What stops an autonomous content agent from publishing something off-brand?

Constraints on the autonomous path itself. Our scheduled publishing job can only pick topics from pre-approved niches and must check that the topic is non-duplicative before drafting. The publishing agent then refuses to ship without verification, refuses to improvise credentials, and refuses to silently skip a platform that failed.

How often can an agent-run content pipeline publish?

Ours runs two publishing cadences — one twice a week and a second client track twice a week on different days — plus a daily QA sweep, a daily publisher health check, and a continuous watcher that moves verified work forward. The cadence limit is topic approval, not drafting capacity.

What is the most common failure in an automated publishing pipeline?

Delivery configuration, by a wide margin. Across our fleet, eight scheduled jobs have hit a state where the work completed correctly and then had nowhere to go — a platform not enabled, or no delivery target resolved. It is invisible in model-level metrics and totally useless to the business, so we treat "delivered nowhere" as a hard failure.

Amit Kumar

Founder of I Am Agent Man. Builds and runs production AI agents on Hermes, OpenClaw, and MCP — self-hosted, model-agnostic, with persistent memory and hard cost ceilings.

Want a publishing pipeline you don't babysit?

Tell us your content workflow. We'll scope the agents, the gates, and the guardrails — free, before any invoice.

Goes straight to hello@iamagentman.com — we read every message ourselves. Prefer to answer three questions instead?Build your blueprint.

Get my agent blueprint — free

No retainer to start · reply within 24 hours