AI content marketing without the AI voice

Searches for this are far higher in the US than in the UK, and most of them are for AI content creation rather than AI copywriting or AI content marketing.

The content agent AI for Marketing built for HEAVENSAKE, a premium Franco-Japanese sake brand, follows that workflow. It was built from the brand's own materials, tone guidelines and approved terminology, it covers research and planning, blog drafts, improving existing landing pages and drafting new ones, and every output is a draft for the brand team to approve. It was delivered in full; HEAVENSAKE chose not to deploy it on their site, so there are no traffic results to report. The build is described in the HEAVENSAKE case study.^[2]

For search, Google's position is clearer than most people think. Its guidance says its systems reward high-quality content however it is produced, and that using automation, including AI, to generate content with the primary purpose of manipulating rankings breaks its spam policies. Its spam policy on scaled content abuse names using generative AI tools to generate many pages without adding value for users as one example. Google puts the rest plainly: "Using AI doesn't give content any special gains. It's just content."^[3][4]

Google's page on using generative AI on your website adds the practical part. It asks you to focus on accuracy, quality and relevance, especially when content is generated automatically, and that includes the title, the meta description, structured data and image alt text, since those can appear in search results too.^[5]

Harmonance, Jakub's own project and not a client, is where the method was first tested. A pipeline researched, wrote and published articles about sound healing with no human review, and Google Search Console recorded 10,976 organic clicks and 770,576 impressions in 152 days with no paid promotion. It shows that AI-written content can earn search traffic. It does not show that unreviewed content is safe for a business with a reputation to protect, which is why client work adds human review before anything is published.^[6]

In advertising, AI copy earns its place as a source of variants. For ToastPal's Google Ads, which Jakub ran as a freelance consultant before he founded the company, AI drafted headlines and descriptions in volume so that new versions could be tested every week, and he decided which ones ran and which were paused on their results. The machine supplied the options; the tests and a person's judgement picked the winners.^[7]

Where AI copy fails is predictable, which makes it easy to check for:

Feed it what only you have: transcripts of sales calls, the questions customers ask, your best past pieces, the numbers from your own campaigns, the words your founder actually uses. A draft written from that material starts closer to your voice, and the editor's job becomes tightening rather than rewriting.

Then edit for the habits that give machine writing away. These are the ones readers notice first:

Google's advice points the same way. It asks publishers to think about who made a piece, how and why; to use accurate bylines where readers would expect them; and to add AI or automation disclosures where someone might reasonably wonder how the content was created. It also says that giving AI an author byline is probably not the best way to make its role clear.^[3][8]

At AI for Marketing, content for clients is written from their own brand material and keyword research, Jakub reviews every piece against it, and nothing is published until the client approves it. The wider question of being cited by AI search engines is covered in how to rank in ChatGPT and Google's AI answers.