Generative AI, competition, companion or completely pointless?

As a web designer and developer, I am always on the lookout for the latest tools and technologies that can help my clients to get the best out of their website. One area generating a lot of buzz rencently is AI-assisted copywriting. Is it the secret weapon to crafting quality website content or will it just produce a load of robotic dribble, pulled from a giant can of well polished phrases?

Let's have a closer look at this undeniably fascinating world of AI-powered writing and explore its potential benefit for your website, it's limitations and possible pitfalls.

The Advantages: Efficiency and Beyond

AI-powered copywriting platforms offer some undeniable benefits:

• Speed: Imagine generating multiple website copy variations in minutes instead of hours. AI tools can churn out content at a rapid pace, saving you valuable time for other crucial tasks.
• SEO Optimisation: Many AI writing tools are equipped with built-in SEO features that analyse keywords and suggest optimised content, boosting your search engine rankings.
• Consistency: Maintaining a consistent brand voice across all website pages can be challenging. AI tools can learn your brand's tone and style, ensuring uniformity in your messaging.
• Overcoming Writer's Block: Stuck on a particularly tricky section of your website copy? AI tools can provide helpful suggestions and spark new ideas, helping you overcome creative roadblocks.

Challenges and Limitations: It's not human after all

While generative AI tools are powerful, they're not without limitations:

• Lack of Originality: AI relies on existing data to generate text, which can sometimes result in generic or uninspired content. True creativity and unique storytelling still require human ingenuity.
• Limited Emotional Intelligence: AI struggles to fully grasp complex emotions and nuances that often resonate with readers on a deeper level. Human writers excel at crafting compelling narratives that evoke feelings and build connections.
• Accuracy and Fact-Checking: While AI can generate grammatically correct text, it's essential to double-check for factual accuracy and potential biases in the output.

Finding the Right Balance: A Collaborative Approach

The key takeaway is that AI-assisted copywriting can be viewed as a valuable tool to support, rather than replace, human creativity. Think of it as a collaboration:

• Use AI for Efficiency: Leverage AI tools for tasks like generating initial drafts, brainstorming ideas, or optimising content for SEO.
• Humanize the Content: Add your own unique voice, perspective, and emotional intelligence to ensure the copy is engaging, authentic, and resonates with your target audience.
• Fact-Check and Edit Carefully: Always review and edit AI-generated text for accuracy, clarity, and consistency with your brand voice.

Two heads are better than one, even if one of them isn't really a head. By embracing this collaborative approach, you can harness the power of AI while retaining the essential human touch that makes your website content come to life.

The Human Touch Remains Vital

In the end, if you want to reach your visitors, the content you put on your website has to come from you, it needs to have your personal touch, your unique signature. If that's missing, people will notice. Throw drafts, headlines or paragraphs back and forth between yourself and the AI, they are really good at brainstorming, but you should always do the final proofreading and literally literally have the last word.

A typical Workflow: This is how I do it

When I do AI assisted writing, my usual workflow goes like this:

1. Setting up the AI

   First I tell the AI, that I need it's assistance in a copywriting task and provide it with some context about the purpose and intent of the article.

2. First draft

   • If I have a rough draft, I input it into the AI and ask for suggestions.
   • If I just have an idea about a topic, but I can't really find a starting point, I explain that general idea to the AI as I would explain it to anyone I wanted to brainstorm with and ask it to write me a first draft.
   In both cases the first edit by the AI is usually very well structured but sounds often a bit too polished or "salesy".

3. Bringing it to life

   The well structured nature of the first revision by the AI makes it an ideal starting point. In this step I read the entire draft from beginning to end and make changes or rewrite parts that have gone to far astray from my personal style, as I read along. Wherever I am not happy with the phrasing, but cannot yet decide how to change it, I make a note.

4. Break-time

   Breaks are important. Particularly in writing, where nothing is worse than getting stuck in a thought. Once you get stuck, the more you turn in circles, the stronger you enforce the synaptic connections in your brain that form that very way of thinking that you got stuck in, making it harder to think out of the box. Better known as writer's block.

5. First review

   After having replenished my energy with a nice cup of coffee or tee, I read through the entire draft again with a fresh mind and make changes where needed. If there are still parts that I cannot decide about, I throw those sentences back at the AI and ask it for alternative suggestions. Once I am happy with that revision, I "show" it to the AI, sometimes with explanations, why I preferred my changes over it's original wording to help it better understand my intent, and finally ask something like: What do you think? And just like a real human, most of the time, the response is not "perfect, go with it!". Most of the time it gets back with a slightly revised version, explaining it's suggested changes in footnotes.

6. Break time

   Yes, again.

7. Second review

   At this point I go through the suggested changes and decide which ones I keep, which I'll discard or edit the bit inspired by the suggestion. Most of the time this goes along with a lively brainstorming session between me and the AI, where I throw it snippets, questions or thoughts about particular bits and get back mostly useful suggestions, alternatives or even general considerations.

8. Proofreading

   Once I am happy with the complete document and just want to get rid of any mistakes, inaccuracies created by the AI and the odd clumsy phrasing, I read through the complete document again, fact-check wherever necessary and only make changes, whenever I find a mistake or a really unfortunate phrasing. Then I feed it back to the AI for the last time, letting it now, that this is my final revision and I just want it to have the grammar and maybe style checked. Usually it this point the AI just confirms that everything is okay.

9. Publication

   Finally I publish the work, usually not acknowledging my co-author in the credits.

This time though I will credit my writing partner: I have developed this very article in a rather productive brainstorming session with my self hosted Ollama AI, running mainly the open source Large Language Model Gemma, developed by Google.