AI is genuinely useful for some marketing jobs and genuinely damaging for others. This service works out which is which for your business, sets up the process, and puts the checks in place that keep the output accurate and worth publishing.
Never let a model be the source. It can organise what you know. The moment it supplies the facts, nobody has checked them.
What generative AI for marketing means
Using language and image models as part of how marketing gets produced: research, structure, first drafts, variations, summarisation and repetitive transformation work. It doesn’t mean pointing a model at a keyword list and publishing what comes out.
Google’s own guidance is clearer on this than most commentary. It says generative AI is useful when researching a topic and to add structure to original content, and that its scaled content policies apply no matter how content is created. AI use is fine. Volume without value is not.
Who it’s for
This suits businesses already using AI, or about to:
- teams publishing more than their people can write properly
- businesses whose AI-assisted content isn’t performing and can’t work out why
- anyone worried about the risk of publishing something wrong
- teams with repetitive production jobs that a person shouldn’t be doing
Where it genuinely helps
Research and structure. Pulling together what’s been published on a topic and organising it is work AI does well and quickly.
First drafts of something you know. A model drafting from your own material, briefed by someone with the expertise, saves real time.
Transformation jobs. Turning a recording into notes, a document into a summary, or a long guide into a set of questions. The source is yours, so the output is grounded.
Variations at scale. Ad copy, meta descriptions and product variants, where the underlying facts come from your data.
Where it doesn’t
Anything requiring first-hand experience. A model can’t have used your product or done your job. That’s exactly the material that makes content worth reading, and it’s the part that has to come from a person.
Facts you can’t check. Models produce plausible, confident, wrong statements, most dangerously with statistics and specifications. Every factual claim needs tracing to a primary source, and that’s the step people skip.
Anything where being wrong is expensive. Regulated advice, safety information, specifications that affect whether a product fits.
Volume for its own sake. Publishing more pages than you can stand behind is the failure mode the spam policies describe.
What’s included
- An honest assessment of where AI would help your marketing and where it would cost you
- A production process with the human steps defined rather than assumed
- Fact-checking standards, including what counts as an acceptable source
- Brief templates that give a model the material it needs to draft from
- A review checklist covering accuracy, voice, originality and disclosure
- A disclosure position, since Google recommends being transparent about how content is created
- Training for your team on the parts that can’t be automated
The rule I’d give you for free
Never let a model be the source. It can help organise what you know and draft from material you supply. The moment it’s supplying the facts, you’re publishing something nobody has verified, and the failure is confident rather than obvious.
What you get
- A written process your team can follow
- Brief and review templates
- A sourcing standard for factual claims
- A clear list of the work that stays fully human
Limitations
AI tools change constantly, so a process built today needs reviewing. Output quality also depends heavily on the material supplied, which means the briefing step is where most of the value sits and where most teams under-invest.
What this won’t solve
AI won’t give you expertise you don’t have. If nobody in the business has done the thing, a model can’t write about it credibly, and the honest answer is to bring in someone who has. It also can’t take responsibility for accuracy, which stays with you.
Talk to the person doing the work
There is no account manager here and no team to be handed to. If you get in touch, it is me who reads it, me who looks at your site, and me who does the work if we go ahead.
Get in touch, or start with a free assessment.
Engagement and price
There are no list prices here, and that is deliberate. What a piece of work costs depends on the state of the site, how much of it there is and what you actually need, none of which I know before looking at it.
So it starts with a free analysis. I look at your site, your Search Console and who you are really competing with, then come back with what I would do first and what it would cost to do. No obligation attached, and if the honest answer is that you do not need me yet, I will say so.
How the work gets done
Content produced this way runs through the same content quality scoring as anything else, with sources checked through the crawl and data integrations. This service sits under AI search visibility. For content produced to this standard, see blog and article writing and content pipelines and editorial briefs.
FAQ
Does Google penalise AI-generated content?
Not for being AI-generated. Google’s position is that its policies target content produced primarily to manipulate rankings rather than to help people, and that those policies apply however the content is made. Poor content is the problem, not the tool used.
Should I say my content is AI-assisted?
Google recommends being transparent about how content is created, and a short statement about your process costs nothing. It also tends to improve the process itself, because a team willing to describe how something was made usually makes it more carefully.
Can AI write my blog posts?
It can draft from material you supply, briefed by someone who knows the subject. What it can’t do is provide the first-hand experience and verified facts that make a post worth reading, so treating it as the writer rather than a drafting tool is where sites get into trouble.
What’s the biggest risk?
Publishing something confidently wrong. Models produce plausible statistics and specifications that don’t exist, and unlike bad writing it isn’t obvious on the page. Tracing every factual claim to a primary source is the step that prevents it, and it’s the step most often skipped.
Work out where AI fits
Get in touch and tell me what you’re currently producing, and I’ll tell you honestly which parts AI should touch.