Will Google Penalise AI-Written Posts? How Revnu Researches and Checks Each Post
Google ranks posts on whether they help the reader, not on who or what wrote them. How Revnu picks topics, researches and fact-checks every post, what it refuses to publish, and how to review each post before it goes live.
Google ranks helpful, accurate posts however they were written, and that includes posts written with AI. Its published guidance says it ranks content on whether it's helpful and reliable, however it was made. What it acts against is pages churned out in bulk to game rankings, with little value for the reader, whether a person or a tool wrote them.
So the real question is whether each post answers a real question accurately. That's also what gets a post cited by ChatGPT, Perplexity, Gemini and Google AI Overviews: clear, specific, sourced answers. See how AI engines choose what to cite. This page covers what Revnu does on every post to make that true, the quality bar each post clears, and how to read every post yourself before it goes live.
What Google says about AI-written content
Three Google Search Central pages cover it:
"Google Search's guidance about AI-generated content" (February 2023). Google rewards high-quality content however it's produced.
"Creating helpful, reliable, people-first content". The questions Google asks of any page: is it accurate, original and written for the reader rather than for search engines.
Spam policies, "Scaled content abuse". Producing many pages mainly to manipulate rankings, with little value, breaks the policy no matter how the pages were made. Generative AI is the first example it gives.
Nobody can promise you a ranking, and Revnu doesn't. What it controls is what goes into each post and how fast posts go out.
How Revnu picks what to write
Revnu writes one post a day on its own, on a topic it can tie to a real signal:
A query Google already shows you for, with no post of yours answering it (needs Search Console)
A buyer question you're missing from in AI answers (see AI visibility)
A keyword you've told Revnu you want to rank for
A topic a competitor's blog covers and yours doesn't
A question a prospect or you asked
Before it starts, Revnu checks the topic against every post you already have. If one already covers it, Revnu picks a narrower angle or a different topic. It won't write a second post on ground you already own.
Every draft carries a short Why This Post note in the editor, naming the signal it came from, so you can judge the topic before you approve it.
How each post is researched and written
It reads your business. Before your first post, Revnu writes Product Facts from your own site: what your product does, its features, integrations and pricing. You can read and edit them in Settings > Memory > Product Facts.
It researches the topic. It reads your business context and your existing posts, then searches the web for the topic and the sources it will cite.
It writes from your facts. Every claim about your product has to come from Product Facts. The writer is told not to add features, integrations or numbers that aren't there.
It edits for plain language. A second pass rewrites the prose to read like a person wrote it, without changing any fact.
How each post is checked before it can publish
Every post goes through these checks before it reaches Review:
Claims about your product are checked against Product Facts and your business sections in Memory. A claim your facts don't support is cut, or softened to what they do support.
Facts about the outside world are pulled out and checked against the web: numbers, named products, dates, prices and anything cited to a named source. Each cited sentence is checked on its own, so an invented study can't hide behind a real one. A wrong fact is corrected with a source. One that can't be stood behind is removed.
Sources must be pages Revnu actually read while researching. Any link that returns "not found" is dropped before publishing.
Never Say is checked three times: when writing, during the claim check, and in a final word-for-word check of the finished post.
The results are in the Fact-check panel of each post in Visibility > Posts.
The quality bar every post clears
Revnu holds back anything that falls short, including:
A post that failed the fact-check. If the check found a claim that contradicted the facts, or a correction didn't apply, the post is marked Failed and doesn't publish on its own. Revnu usually fixes it or writes a replacement before it asks you anything. If a Failed post sits there, ask Revnu to rewrite it.
A post that names anything on your Never Say list. It's stopped the same way, and Revnu rewrites the wording before it asks you to approve it.
Made-up proof. No invented customers, quotes, hours saved, ROI, rankings or market share. A customer named in your account isn't permission to use them in a post.
Ranked lists that crown you. Comparison posts are written from each competitor's current public documentation and compare how things work. Revnu doesn't publish "top 10" lists with invented winners or rankings that put you first.
Private sources. Nothing from your chats, calls or internal tickets is cited.
More than it can get indexed. One post a day on its own is the ceiling. With Search Console connected, Revnu checks its recent posts are in Google before writing more and holds off when they aren't. You get time to read each post, and Google gets time to crawl it.
A claim Revnu simply couldn't confirm, because no public source covers it, doesn't stop a post on its own. If a held post has claims like that, Revnu points them out so you can check them before approving.
Teach Revnu your facts and your rules
Two sections in Settings > Memory steer every post:
Product Facts is the source for every claim about your product. Fix a wrong fact here once and every later post uses the correction. Short, literal lines work best, for example "Minimum plan: 5 seats" or "We no longer sell the desktop app".
Never Say is a list of things that must never appear in a post: a claim, a word, a retired product, a competitor's name. One per line.
Or tell Revnu in chat, and it records the fact and fixes the live posts that carry the old claim:
We don't offer a free plan any more. Remember that, and fix any live post that says we do.
Control what gets published covers editing live posts.
Read every post before it goes live
On Pro, Blog Posts is set to Just Do It by default. A finished post publishes at 9:00 on the first weekday at least 24 hours after it's ready, unless you act. To read every post first:
Open your approval settings
Open Settings and choose the Agent tab. Find What Needs Your Approval.
Set Blog Posts to Ask Me First
Switch the Blog Posts row off, so it reads Ask Me First.
Approve each post when it's ready
Open the post and click Approve & Publish. Nothing publishes until you do.
Drafts already waiting lose their publish time. Revnu still writes a post a day for you to approve. The same setting covers Revnu's edits to posts that are already live. Switch it back to Just Do It later, once you trust what it writes.
Tip
For a post about something legal, medical or financial, keep Ask Me First on and read it yourself.
Next step
Frequently Asked Questions
Not for being written with AI. Google's guidance says it ranks content on whether it's helpful and reliable, however it's made. Its spam policies target pages mass-produced to game rankings with little value. Revnu writes one researched, fact-checked post a day, tied to a real search or buyer question.