Bringing Retrieval-Augmented Content Generation (RAG) into the Magnolia AI Accelerator
Juni 3, 2024
--
RAG 1200x628

Bringing Retrieval-Augmented Content Generation (RAG) into the Magnolia AI Accelerator

The IT world started to move considerably faster in November 2022. This was the date when GPT3 became available to the general public. A lot of things have happened since then, including the advent of our AI Accelerator program.

As the landscape evolved, so too did our ambitions, pushing us beyond initial experiments and plans to deliver robust solutions that significantly enhance content creation. Since quite a while now, we're no longer just playing with cool features, we see more and more use cases that really move the needle for our clients.

Our latest release will push the boundaries again. Time to have a look at some of the key innovations.

Introducing Retrieval-Augmented Content Generation (RAG)

One of the most significant advancements is the introduction of retrieval-augmented content generation (RAG), which allows you to tap into various data sources and incorporate that information into your core hyper prompts. This means that you’re enhancing the reliability and accuracy of the content generated by the AI models you’re using with facts fetched from external sources - typically the core business systems you use to store data about your products, customers, or processes.

This is a game-changer because it opens up new possibilities for content creation based on your core business data.

Let's make it concrete based on an the task of producing a product story with the help of PIM (Product Information Management) data:

Imagine you're tasked with creating a story about an engineer who crafted a pair of premium headphones, for simplicity we will call them the H66. With RAG, you can pull data from your other sources to enrich your prompts​. This means you can generate content that's not only creative, but also accurate and detailed.

Generating content with RAG: A story about the H66 headphone

Let's generate a story about an engineer and his journey in designing the H66 headphones. With RAG, we can include specific product names, technical specifications, and design details in the prompts. This creates content that's not only engaging but also deeply connected to the product's unique attributes​.

RAG 1

Here's a snippet from the generated story:

“As the lead product designer at Prestigio, Joe Worwich was tasked with creating the ultimate over-ear headphone that would revolutionize the way people experience music. With a passion for audio engineering and a keen eye for design, Joe set out to create the H66 Over Ear Headphone - a masterpiece that would redefine the standard for premium audio products.

The first step in the design process was to select the finest materials that would not only enhance the sound quality but also provide a luxurious feel for the wearer. After extensive research and testing. Joe decided to use a combination of aluminum, titanium, lambskin, cowhide, and polymer textile. These materials were chosen for their durability, comfort, and ability to enhance the sound experience.

Next, Joe focused on the audio functions of the H66. With advanced features such as Adaptive Active Noise Cancellation and Transparency mode, the H66 was designed to provide pure sound and tranquility to the wearer. The titanium electrodynamic driver with a 40 mm diameter ensured a wide frequency range 20-22,000 Hz, allowing for a rich and immersive audio experience.

One of the key features of the H66 is the individually customizable sound EQ. Users have the option to choose from presets available on the Prestigio app or fully customize their sound profile to suit their personal preferences. This level of customization ensures that each user can tailor their audio experience to perfection.

The design of the H66 was carefully crafted to not only look visually stunning but also to provide optimal comfort for long listening sessions. The over-ear fit and adjustable headband ensure a perfect fit for any wearer, while the IP53 rating protects against dust and water, making the H66 suitable for use in any environment.

With dimensions of 185 W x 165 H x 80 D for the headphones and 200 W x 180 H x 63 D for the aluminum housing, the H66 is both sleek and stylish. Despite its premium quality, the headphones weigh only 323g, while the aluminum housing adds an additional 500g.”

If you’re a copywriter or content creator, here’s a demo where we show how you can use all that in the Magnolia editorial workflow:

And if you're a designer or in some other creative role, the same capabilities are available in the Magnolia DAM. This demo video shows what you can achieve with AI when it comes to generating and managing images and other artwork:

Now that you’ve seen how powerful RAG is, let's have a look on a small selection of opportunities that utilizing RAG offers:

Product-specific stories

Creating detailed stories about the design, engineering, or manufacturing process of your products can engage customers and build brand loyalty. These stories can delve into the unique aspects of a product, providing insights into its development and showcasing the team's creativity and innovation.

Comprehensive product guides

Detailed product guides and FAQs can provide valuable information to customers, helping them make informed purchasing decisions. AI can automate the creation of these guides, including instructions, troubleshooting tips, and other relevant information, making it easier to produce comprehensive product documentation.

Custom landing pages

Creating custom landing pages for marketing campaigns or new product launches can be time-consuming. AI can automate the creation of these pages, allowing for rapid generation and customization based on specific campaigns or product releases.

Automated content personalization

Personalize content to fit the preferences of your customer segments. This can include tailoring articles, components or newsletters.

Content optimization for search engines (SEO)

AI can be a real game-changer when it comes to SEO keyword optimization. You can effortlessly identify and integrate high-performing keywords into your content, making sure your articles and web pages are more likely to be seen by your target audience once new trends arise. It's like having a super-smart writing assistant which will help boost your visibility in search engine rankings. This way, you can focus more on creating great content while AI takes care of making it discoverable.

Customization and flexibility: the magic of plugins

What's remarkable about Magnolia's AI Accelerator is its flexibility. You can customize the content generation process with your own plugins, allowing you to extend, rewrite, or reshape the content as needed. Whether you want to focus on specific prompt-based steps or integrate data from your data sources, each field in Magnolia's AI dialogs can be customized to fit your content needs.

Let's have a look at a very simple example. By exchanging standard Magnolia fields to AI fields you get full access to AI, including your own workflow-specific text fields.

RAG 2
RAG 3

So as you can see, with a single click basic ideas can be transformed into exceptional content in seconds. And so can content optimization for certain keywords be applied.

In this final demo video,  we show how you can customize the editorial process with your own plugins to leverage AI at every step of the content creation in a highly flexible way.

Wrapping it up: a new era of content creation

Magnolia's AI Accelerator has truly ushered in a new era of content creation. With retrieval-augmented content generation, hyper prompts, and customizable plugins, you can produce high-quality content at amazing speed. Whether you're crafting stories about your products, generating images for your campaigns or building entire websites, Magnolia's AI Accelerator offers the tools and flexibility you need.

If you're interested in learning more about these features, we encourage you to explore Magnolia's AI Accelerator and see how it can transform your content creation process. Trust us, once you experience its capabilities, you'll never look at content creation the same way again.

Final thoughts

The age of AGI (Artificial General Intelligence) is not yet here. Therefore one can't just run an AI service and expect mind blowing results from the get go. The true magic happens once clients and partners look closely at their business domain and apply design thinking to the content bits that couldn't be done because of time or resource constraints.

Various tasks like the described product stories can be very labor extensive and might not be cost efficient at scale. With RAG they are. Let's say you want to localize and personalize the content for your different outlets… Same story: With the right creative thinking suddenly a lot of things that were out of reach before can now be done. So right now we are at a stage in the AI journey where a lot of creativity and experimentation can lead to magical and transformative outcomes. And that is a wonderful thing.

To benefit from the RAG capabilities, get the latest version of the Magnolia AI Accelerator.

Über den autoren

Tobias Kerschbaum

Solution Architect, Magnolia

Als Solution Architect arbeitet Tobias eng mit Kunden und Partnern zusammen und teilt sein Wissen und seine Erfahrung. Er hilft Unternehmen dabei, zu bewerten und zu verstehen, wie Magnolia die Projektanforderungen erfüllen kann. Er trägt zum Projektplan bei und stellt sicher, dass die richtigen Module und Technologien ausgewählt werden. Neben der Durchführung von maßgeschneiderten Workshops wird Tobias auch aktiv, wenn Kunden und Partner neue Funktionen oder individuelle Anforderungen implementieren müssen.

Jan Schulte

Head of Group Consulting, Magnolia

Durch seine Arbeit an der Schnittstelle zwischen Vertrieb und Technologie hilft Jan den Kunden von Magnolia, ihre Initiativen zu Content Management und Digital Experience zu meistern, indem er Lösungen entwirft, die ihren individuellen Herausforderungen und Möglichkeiten entsprechen.