Publishing reimagined – harnessing the power of generative AI in content creation

Written by Magdalena Ivanova | Aug 27, 2024 6:38:39 AM

Going beyond theory

How do you prepare for a journey when you don't know what to expect?

What will you need along the way and what scenarios do you consider upfront?

Tom: “These were a couple of questions that we asked ourselves and the audience of the WoodWing Uncovered event last month in Amsterdam. Particularly considering the fact that the possibilities of AI are seemingly endless. As one of the major software suppliers for the publishing business, we are even more eager than the average person or organization to find answers ahead of time.”

“That's why at WoodWing, we run an ‘AI in publishing’-experiment every Friday to see if we can take a couple of steps in our journey to crack the AI-enigma”, Dani continues. “We start each session with a question. For example ‘Can AI help me to copyfit content?’ We discuss the problem and then break down in subgroups to validate and document various ways of solving the case. At the end of the day, we discuss the results and decide on the next steps as a team.”

Tom: “Not every company has the resources to dedicate to this kind of activity. That's why we would like to share some of our latest experiments to see how we can reimagine publishing with the help of AI in content creation.”

Can AI help me to transform articles?

The solutions we are investigating keep the creatives in control. The artists are the superheroes and AI is the sidekick. It's a good balance of how the two fit together.

We looked at conversion of print to digital, translations, SEO optimization, creating summaries, conversion of existing content to social content, and text analysis and tips. This is what we learned:

Negative	Neutral	Positive
Margin for error in transformation is low	Interest in transformation depends on how articles are transformed Brands want to know how rich the text is in transformation	Transforming articles with AI is easier than the existing rule-based conversion Brands are focused on refreshing evergreen content. If this transformation could slightly alter the existing library of evergreen content, this could be big

Can AI help me to find related content?

The hardest challenge in generative AI is getting the writers to use the tools. People are scared. But it's not about replacing anyone, rather like saving a tremendous amount of time. What’s interesting here, is that everything you’ve shown shortens the cycle. You could move everything lightspeeds faster.

Can we create something in the vein of Google Search, find related information and use this information to create new content? This is what we learned:

Negative	Neutral	Positive
Many people are nervous about AI and large language models. If users are told they will be using an LLM, they are concerned that technology will be replacing their jobs. Instead, we want to leverage the power of ground-breaking technology in a way that is natural, intuitive, and a lot less scary to users.	The ability to filter on rights and usage is needed	Combining searching articles with transforming articles is a compelling use case This idea supports the emphasis on using all available content for topicals, including previous magazine content and recent website additions This feels like a new way to drill down into your content

Can AI help me to copyfit content?

It's a known fact that print is under pressure. Further efficiency is absolutely required. One of the time-consuming tasks is adapting the text so it fits the design. The question is, how can we do this without losing creative freedom and whilst ensuring humans remain in control? This is what we learned:

Negative	Neutral	Positive
AI-generated text lacks emotion Text styling is not considered Short sentences should not be shortened further	Many different workflows: print-first and digital-first Users copyfit from top to bottom Magazine style content is copy fitted in the middle Copyfitting is a complex task for inexperienced users	Users like the fact that they remain in control Useful in case of writers block Useful for larger articles Manual and AI copy-fitted text are rated equal by control group

@WoodWing, we are breaking down barriers with generative AI

By experimenting together with customers, we see AI’s potential for growth every day. When customers see the solutions we’re working on, they cannot help but get excited. Working toward the goal of increasing efficiency helps creatives understand how they could work with a complex tool like Generative AI.

We overcome unfamiliarity with AI in several ways:
- slowly introducing editors helps them understand the possibility
- Everything is experimental and creative rather than an impending shadow
We know that keeping an artist or writer in control is mandatory.
We are very strict when it comes to ethical considerations:
- the data we work with is either publicly available, donated to us, or under NDA
- at all times, we respect the customer's data and how they want it to be treated

WoodWing and AI

WoodWing has been active in mapping the possibilities of AI in different areas other than AI in publishing and AI in content creation. If you're interested in finding out more about WoodWing's look on AI in general and specifically on AI in content creation, and the way it can help you move forward in different markets and areas of expertise, make sure to check out the following articles:

View full post