How Did We Increase Our Conversion Rates by 566% with GenAI?

Searching for Generative AI in the World’s Biggest Dump

Placebo’s video for Life’s What You Make It, released six years ago, is dedicated to the recycling workers of Agbogbloshie, a commercial district near the center of Accra, the capital of Ghana, which sits on the Korle Lagoon of the Odaw River and owes its fame to being the world’s largest electronic waste dump.

The song draws its magic from the idea that, even inside the world’s largest electronic dump, life is shaped by what a person does: Life’s What You Make It.

So is generative AI.

With the launch of ChatGPT, which can be described as the biggest leap in the “consumerization” of information technology since the arrival of the iPhone in 2007, the way we work, our productivity and our outlook on life changed and deepened like never before. Untouched thoughts waiting to be discovered in the farthest folds of our minds met the sun and came to light. Creative ideas waiting to be united across the farthest distances of the world, every inspiration we had written off as impossible, became reality.

54 years after the footprints Neil Armstrong left on the moon at 05:56 Turkish time, this time a mark that will find its place in humanity’s future was left on Earth: ChatGPT. And so the first drop of the generative AI downpour fell.

Although at its core this whole story heralds a change unlike anything seen before, my aim is never to sing the praises of generative AI. After all, GenAI still holds dozens of challenges waiting to be discovered and solved, and understanding its potential and use cases and building a model that actually works is still quite hard.

So how will we find our own way?
PS: We found it.

Last spring, under the roof of Komtas.com, we launched Information HUB, Turkey’s largest data and analytics library. Made up of 5 main sections, Information HUB is a living library that features up-to-date content for the ever-evolving world of data, with its new terms and trends, and contributes to sustainable success.

While launching this library, we struggled a lot with one thing: for the dozens of e-books, whitepapers and use cases in Information HUB, we wanted to

• extract their essence,
• show the visitors who would consume this content a meaningful summary of it,
• and offer meaningful copy on what they would gain by reading them.

Bringing the right content together with the right person in the right form, and presenting it consistently and persuasively across hundreds of documents.

When there’s an elephant in the room, you have to talk about the elephant first. Talking about the potential of generative AI is of course meaningless, but the world we define today through the outputs of generative AI is no different from Agbogbloshie, one of the world’s largest electronic dumps, and as Placebo says, the life (future) we will build for ourselves in this dump is directly tied to how we use GenAI.

In this context, through days of prompt preparation, trial and error, and the thrill of building a working model, we managed to build a model that works consistently across dozens of documents, produces persuasive output and, most importantly, can deliver in one hour the content an employee could produce by working on it for months.

With this model:

• we increased our conversion rate by 566%,
• we strengthened our lead generation processes,
• with more consistent content, we raised the quality score of our ads.

Next up is a production line and a product we will build on top of this model. Let’s see which geographies this new world, these undiscovered lands, will introduce us to in the hands of modern explorers.

MH.

Follow me on LinkedIn: linkedin.com/in/ahmetsenkaya