top of page
  • Writer's pictureZesta

Content Marketing in a Generative-AI era

Updated: Oct 13, 2023

Generative AI is rapidly transforming the world around us, and content marketing is no exception. AI-powered tools can now help content marketers to create, distribute, and optimise content in a fraction of the time it would take using traditional methods. On the other hand and according to Salesforce's State of the Connected Customer report, 68% of consumers say advances in AI make trust even more important, but only 51% say they trust companies overall.

What does this mean for content marketers in the long term? How can we ensure that our content remains relevant and engaging in a world where AI is generating more and more content all the time? We asked some content specialists for their insights.

A man using chat GPT on his laptop

According to Dominique Cozzi, former journalist and Content Strategy expert, "businesses should not give in to a process where AI would be the sole driver. ". She thinks that "AI should not be seen as an end game but rather as a means to quickly create a library of sources on billions of web pages, to shorten research time, or generate summaries." On the other hand, "content must continue to resemble a living organism with testimonials, feedbacks and interviews. This will give content a human touch that will reassure this audience"

Morten Leth Jacobsen, senior consultant and content creator at Gammelbys in Denmark, has this take on AI: “As a traditional copywriter and as a published fiction writer, my initial reaction was to change careers and become a yak herder in Tibet instead. But I find solace in three things: AI only ever reshuffles stuff that’s already there, it’s not truly a creative tool - you need humans for true creativity, and you need humans to go out and dig up stories. Secondly, AI is a great tool to help you delevop and expand on topics - as long as you take the time to fact-check its findings and use your human critical sense. And third, AI does not really save time. It needs prompting, and its texts need refining by human hands. We use it openly and tell our clients about it when we do. But we try to vaccinate them against the idea that it is a shortcut to creativity - it isn’t.”

Grégory Dominé, Global Head of SEO at L'Oréal, thinks that "the ability of Gen AI - more precisely the Large Language Models (LLM) it’s built upon - to understand natural language in context, provides infinite possibilities to plan comprehensive content strategies that engage customers better. From content plan to content calendar, content segmentation to persona generation and extensive keyword research, LLM are a fantastic tool to augment marketing strategies in the era of big data."

In light of our discussions and based on our experience, we created a short summary of the questions you should ask yourself before embarking on the AI journey:

1 - Should I even think about AI? Yes, yes and yes. Chances are that you will invariably improve your productivity. Also, the objective is not to go full AI. There is no need to buy the latest fast car if you haven't passed your licence yet. Chances are that you're already using AI via your tech stack. All major tech companies are packed with AI already.

2 - Which generative AI should I work with? Hard to answer. ChatGPT, Bard... But also specialised and niche ones such as Synthesia for videos or DALL-E 2 for images. For more advanced users, you can also pick AI Machine Learning tools, such as Vertex AI or SageMaker. Overall, we're afraid that you may have to test them on a regular basis to see how they evolve and what fits your needs.

3 - What can I start with? Depending on what you want to produce content for (e.g. comms, ecommerce, social media?), we would recommend the following steps:

  • AI for ideation: prompts like "how best to use gen AI in order to help companies create content" helped us create this article. The title and 98% of the intro in this article have been created by AI.

  • Content creation: some content are easier to create with AI than others. For instance, product pages on a website or social media posts (make sure you include which tone you want to use). You might also want to consider emails that you send every year to your clients. AI will help you add a different touch each time.

  • Refresh content: AI can entirely refresh a page in order to rank better in Search Engine Result Pages. Best is to pick the content that are no longer performing well and rejuvenate them with new stats, new trends etc.

  • A/B, MVT, personalisation: AI can help you generate various content just by changing one variable. This will greatly help speed up the process of testing.

  • Classification: AI can help classify content into a pre-defined taxonomy or generate one for you. It is quite handy when it comes to Search Marketing or Customer Service.

  • Please double check anything you create with generative AI. Some tools can generate factual errors.

4- What should I do next? Differenciation: same prompts with the same tool connected to the same data may result in the same content being produced. Differentiation will require the ability to fine-tune GenAI models with large volumes of high-quality, firm-specific data. In order to connect your generative AI tool with your database, you may want to consider:

  • Look at the default existing connections. For instance, Bard can connect to Google Spreadsheet.

  • Look at plug-ins: Chat GPT connects to Zappier which can then connect to pretty much everything you need.

  • There are of course APIs like for any other tools :)

  • Start exploring Machine Learning.

5- What else should I know?

Our conclusion is that every business should consider testing out AI technologies in order to improve their productivity and / or scale the production of simple content. On the other hand, your customers will detect both content spin and flavourless content and (human) content specialists are more needed than ever in order to bring creativity, differenciation and authenticity within your content strategy.


bottom of page