What an AI Content Services engagement with us includes
What you get
- AI-assisted content production with human editing on every piece
- Brand voice guide the models are constrained to
- Long-form articles, product copy and landing pages
- Bulk product descriptions and category copy for large catalogues
- Repurposing one asset into social, email and short-form scripts
- Fact-checking and originality pass before publication
How we run it
- Build a brand voice and terminology guide
- Research topics and search intent
- Draft with AI against that guide
- Human edit, fact-check and originality pass
- Publish and track performance
Tools and stack
- Claude
- GPT
- Gemini
- Brand voice prompts
- Human editorial review
- Typical timeline
- First batch typically 2–4 weeks, then a monthly cadence
- How we price it
- Per-batch or monthly retainer by volume and depth
Key capabilities
- A brand voice guide the model is held to
- Built from your real positioning, terminology and the things you refuse to say — so drafts start in your voice, not a generic one.
- Human editing on every piece
- Nothing publishes unedited. The edit is where your data, your customers and your point of view get added.
- Fact-checked and originality-passed
- Claims are verified and every piece is checked for originality before delivery.
- Built for search intent
- Topics are chosen from what people actually search and what your site can realistically rank for.
- Long-form articles
- Researched pieces built around a keyword and a genuine angle, not a word count.
- Bulk product copy
- Thousands of product and category descriptions generated from your attribute data.
- Landing page copy
- Page copy written to a single conversion goal, matched to the ad or query that brought the visitor.
- Repurposing
- One asset cut into social posts, email sequences and short-form video scripts.
- Brand voice guide
- A reusable document your team and ours can both write against.
- Editorial calendar
- A publishing cadence mapped to intent, so the content compounds instead of scattering.
Does AI content get penalised by Google?
Google's position is about quality, not authorship. Its spam policy targets content produced at scale for the primary purpose of manipulating rankings — which describes unedited, undifferentiated output, whether a machine or a content farm produced it. Helpful, original, well-edited content is judged on its merits regardless of how the first draft was written.
That is precisely why the edit is not optional in how we work. The failure mode we are guarding against is publishing a hundred pieces that any competitor could have generated from the same prompt. The brand voice guide and the human pass are what stop that.
Where AI drafting genuinely helps
Bulk catalogue copy is the strongest case, because the source facts are already structured — you have the attributes, and the job is expressing them consistently across thousands of rows. Localisation and repurposing are close behind: turning one researched article into a social sequence and an email is mechanical work that a model does well under supervision.
It helps least where the value is a genuine point of view — original research, opinion pieces, and anything drawing on experience the model does not have. There we use AI for structure and let a human write the substance.
Frequently asked questions
- Will AI-written content get my site penalised?
- Google penalises content produced at scale to game rankings, not the tool used to write it. The failure mode is unedited, undifferentiated output — so every piece we deliver goes through human editing, fact-checking and an originality pass before it publishes.
- How do you keep it from sounding generic?
- We build a brand voice guide with real examples of your positioning, terminology and things you refuse to say, and constrain the model to it. The human edit then adds the specifics — your data, your customers, your point of view — that a model cannot know.
- Can you handle thousands of product descriptions?
- Yes. Bulk catalogue copy is one of the strongest uses for this, because the source facts are structured. We generate against your attribute data and spot-check a sample rather than eyeballing every row.
- Will the content be original?
- Yes. Every piece goes through an originality check before delivery, and the human edit adds specifics — your data, your examples, your positioning — that no model could produce on its own.
- Do you disclose that AI was used?
- That is your call as the publisher, and we will follow your policy. Google does not require AI disclosure for ranking purposes; some industries and some audiences expect it.
- How much volume can you produce?
- Considerably more than a fully manual team, but we scope it against what your site can realistically get indexed. Publishing far past that point creates the thin-content problem rather than solving it.