AI and writing: strategic allies, never rivals

What AI really changes for writers

Generative AI applied to writing refers to the use of language models (such as ChatGPT, Claude or Gemini) to assist professionals in producing written content: articles, web pages, newsletters, reports.

Contrary to what some fear, AI does not replace the writer; it changes the role. The writer becomes an augmented editorial director, capable of producing more, faster, while keeping control of quality and tone.

In French-speaking Switzerland, where the market demands bilingual, culturally adapted and often technical content, this complementarity is particularly relevant. It fits into a broader movement where hybrid skills are becoming essential to remain competitive.

What AI does better than us

Brainstorming and ideation

A writer facing a blank page can spend 30 minutes looking for an angle. AI generates 20 headline ideas, 10 angles and 5 article structures in seconds. The writer chooses, refines and enriches. The time saving is considerable.

Rewording and adaptation

Adapting the same content for LinkedIn, a blog, a newsletter and a press release used to take half a day. AI proposes variations in a few minutes. The writer adjusts the tone and checks consistency.

SEO and GEO volume production

For search marketing on AI engines (GEO), you need to produce structured, factual, semantically rich content. AI excels at generating optimised first drafts that the writer then refines with their domain expertise.

Summarising and document research

Summarising a 50-page report, extracting key data from a study, comparing sources: AI carries out these tasks in seconds with growing reliability. For a complete overview of available tools, see our top 5 AI tools for writers.

What AI cannot do

Authentic storytelling

Telling a story that touches, that creates emotion, that sticks: this remains the exclusive territory of humans. AI produces correct texts, but rarely memorable ones. Storytelling requires lived experience, empathy and an intimate understanding of the audience.

Humour and second-degree

AI struggles with wordplay, irony and local cultural references. A French-Swiss writer knows when to slip in a wink to the local audience. AI remains literal.

Cultural and regulatory nuances

The Swiss market has its codes. Systematic formal address in B2B, linguistic sensitivities between cantons, specific regulations (the FADP, for example): these are subtleties only a writer rooted in the local context masters.

Editorial stance

An opinion piece, an executive's column, a sharp viewpoint: AI tends towards consensus and neutrality. A brand's voice, by contrast, must take a stand, assert and sometimes provoke.

The figures: ROI of AI-assisted writing

The data confirms what practice suggests. According to an MIT study (2023), writers using generative AI cut their production time by 37% on average, with quality perceived as equivalent or superior by readers.

Other indicators deserve attention:

• Productivity: an AI-assisted writer produces an average of 3 to 5 optimised articles per day, against 1 to 2 without assistance. For a Swiss SME publishing 4 articles per month, this means moving to 10-12 without recruiting.
• Cost per piece of content: the average cost of a professional blog article in Switzerland runs from CHF 400.– to CHF 1'200.–. With a well-honed AI-writing workflow, this cost drops by 40 to 60%, as raw writing time falls sharply.
• SEO/GEO: companies that publish regularly (8+ articles/month) increase their organic traffic by 55% on average over 12 months (HubSpot, 2024). AI makes this publishing rhythm accessible to medium-sized organisations.
• Error rate: contrary to received wisdom, a structured workflow (AI + human review) reduces factual errors by 23% compared with 100% human writing under deadline pressure (Stanford HAI, 2024).

ROI is therefore measured on three axes: production volume, reduction in unit costs and improvement in organic visibility.

Concrete workflow: the human + AI process in 6 phases

Beyond the 4-step method described below, here is a complete operational workflow we deploy with our clients.

Phase 1: editorial planning (human)

The editorial lead defines the publication calendar, priority topics and target keywords. This phase rests on knowledge of the market, customers and the company's strategy. AI does not intervene here.

Phase 2: research and enriched brief (human + AI)

The writer formulates a detailed brief. AI is called on to complement document research: recent studies, statistics, competitive analysis on the topic. The writer validates and sorts the relevant information.

Phase 3: first draft (AI)

Based on the brief, AI generates a first structured version. This draft serves as the skeleton: titles, subtitles, main arguments, key data.

Phase 4: editorial enrichment (human)

The writer takes the draft and injects what AI cannot provide: concrete examples drawn from client experience, brand tone, storytelling, local cultural nuances, stances. This is the phase that turns generic text into high value-added content.

Phase 5: SEO/GEO optimisation (human + AI)

AI helps check semantic density, tag structure, presence of structured data. The writer ensures the content meets GEO criteria: direct answers to questions, source citations, factual structure.

Phase 6: final quality control (human)

Complete review: fact-checking, tone consistency, conformity with the editorial guidelines, internal link control. No content should be published without this step.

Quality control: avoiding AI's pitfalls

Generative AI commits specific errors that you must learn to spot.

Factual hallucinations. AI sometimes invents statistics, quotes or references. Every figure must be verified at the source. This is non-negotiable.

Generic tone. Without precise instructions, AI produces a uniform, recognisable style: short sentences, systematic bullet lists, no personality. The writer must break this pattern and impose the brand voice.

Semantic redundancy. AI tends to reword the same idea from several angles without adding new value. A good editor cuts these passages without hesitation.

Regulatory compliance. In Switzerland, the new Federal Act on Data Protection (nFADP) imposes specific requirements on marketing content. AI does not know these constraints. The writer must check them systematically.

Best practices: the 4-step method

1. Write a structured brief

AI is only as good as the brief you give it. Specify: target audience, desired tone, key messages, length, SEO/GEO constraints. A 10-line brief produces a result 10 times better than a vague prompt.

2. Use AI for the first draft

Let AI generate a first version. Do not seek perfection: seek structure, ideas, flow. It is a starting point, not a destination.

3. Always review and enrich with humans

Every AI-generated text must go through an in-depth human review. Check the facts, adjust the tone, add concrete examples, integrate your expertise. That is where value is created.

4. Train your teams in effective prompts

The quality of the result depends on the quality of the prompt. Training your employees in prompting techniques is a profitable investment. A writer who masters prompting triples their productivity.

The stakes for Swiss companies

In French-speaking Switzerland, companies that integrate AI into their content production gain a measurable competitive advantage. They publish more regularly, cover more topics and improve their visibility on AI engines.

But this advantage only lasts if humans remain at the centre of the process. 100% AI-generated content, without review or enrichment, is spotted and loses credibility.

At MCVA Consulting, we support Swiss companies in this integration: from team training to setting up AI-writing workflows optimised for GEO.

Implications for Swiss businesses

The Swiss economic fabric has characteristics that make AI writing adoption particularly strategic. This dynamic is also visible in creative agencies, where AI is reshuffling the cards of creation.

Multilingualism. A company active in several language regions must produce content in French, German, Italian and even English. AI considerably accelerates linguistic adaptation, even if a native review remains essential for each language. Time saved on localisation is estimated at 50-70%.

SMEs and limited resources. The majority of Swiss companies are SMEs. They have neither the budgets nor the teams to maintain a high publishing rhythm. AI allows a 1- to 2-person team to produce a content volume comparable to that of a 4- to 5-writer team.

Regulated sectors. Banking, pharma, medtech, watchmaking: these key sectors of the Swiss economy demand strict editorial rigour. AI provides a solid base, but human sectoral expertise is the safeguard. Companies that master this balance stand out.

GEO and AI engine visibility. AI search engines (ChatGPT, Perplexity, Google AI Overviews) favour structured, factual and regularly updated content. AI writing makes it possible to maintain this update rhythm. Swiss companies investing in GEO now build a measurable lead over their competitors.

FAQ

Will AI replace professional writers?

No. AI replaces repetitive, low value-added tasks: document research, first drafts, rewording. The writer focuses on what makes the difference: editorial strategy, storytelling, brand tone, sectoral expertise. The role evolves; it does not vanish. Writers who master AI grow their market value.

Does Google penalise AI-generated content?

Google does not penalise the use of AI in itself. Its official policy (March 2024) is clear: what counts is content quality, not the production method. AI content enriched by an expert, fact-checked and useful to users will be well ranked. By contrast, AI content published without review, generic and without added value, will indeed be downgraded, like any low-quality content.

What budget should I plan to integrate AI into a writing workflow?

AI writing tools cost between CHF 20.– and CHF 200.– per month depending on feature level. The real investment is in team training (one or two days is enough) and workflow setup. For a Swiss SME, count on an initial budget of CHF 2'000.– to CHF 5'000.– for training and setup, with ROI visible from the second month thanks to productivity gains.

Operational summary

• AI is an accelerator for writing: brainstorming, rewording, SEO/GEO volume.
• It does not replace storytelling, humour, cultural nuances or editorial stances.
• The winning method: structured brief, AI draft, human review, prompt training.
• Swiss companies that combine AI and writing expertise publish more and better.
• Contact MCVA Consulting to set up an AI-writing workflow tailored to your company.

Want to integrate AI into your content production without sacrificing quality? At MCVA Consulting, we train your teams and set up tailor-made AI-writing workflows suited to the Swiss market. Let's talk.