Start-up: Text Cortex AI
The founder of Text Cortex AI: Ceyhun Derinbogaz and EBS alumnus Dominik Lambersy
How did the idea for Text Cortex Al come about?
My co-founder, Ceyhun, had the initial idea. He is the technical genius of us both. He has been working on the topic of artificial intelligence and natural language generation for several years now. I have spent the last few years developing the business side of AI and Natural Language, first during my studies at Imperial College London and then at venture capital fund Speedinvest.
What once started as a small project around creating domains for Ceyhun turned into a thirst for knowledge. At the beginning of 2021, he began generating AI-driven product descriptions, thus arousing initial user interest in the topic. A short time later, we met at the well-known talent investor Entrepreneur First.
I was immediately hooked by the enormous potential this project would have. Humanity produces more than 8 billion US dollars worth of text content in blogs and social media articles every day. And this is still a conservative figure when you consider that every website and every online shop requires vast amounts of text. Because without text, you don’t appear in the search results of search engines in the digital world - and are therefore not visible.
A small group of leaders in content marketing see 7.8 times more visitors to their website and experience a six times higher conversion rate thanks to the positive brand effect. But many cannot cope with the bulk of the content.
The problem is that human beings write all this content in a linear production for an exponentially evolving digital world. A demand that is either not met at all, too expensive, slow or far too difficult for many agencies and individual writers to cope with or scale. This is where our technology comes in, automating the creation process as we know it.
How does your service function?
Everyone knows Google Translate, DeepL or Grammarly and the time these translation and correction technologies save. Well, we start one step before that. To translate something, a person must have already gone through the creative process and created something. It is precisely in this process that our product helps. It writes text content based on minimal input from the user and produces a unique text with each click. We keep in close contact with our users and have discovered that some save up to ⅔ of their time.
In a very abstract sense, we build tiny, specialised AI helpers tailored to the needs of our users. We train them in different text types, topic areas, and soon even in our users’ corporate identity and writing style – packaged as a typical software service product. Via our web app, you can begin generating text content in less than five minutes. Our little helpers need just two things to start the process:
You have to tell them what kind of text content you require: A blog item, a product description, an attractively worded email. Because our models have a basic understanding of human language, they can even generate text as specific as a piece of advice on venture capital (this was the first stress test I tested with Ceyhun’s MVP back then).
Our AI also needs a title, some context or a few keywords to start in a topic area. With these two pieces of information, our scribes can then get started. In less than 30 seconds, our AI can create a text of several hundred words qualitatively indistinguishable from that of a human.
Of course, not every generated content will fit. But it makes a difference whether you create and formulate a blog article with around 2,000 words from your thought process or have your initial thoughts embellished by AI.
What are the biggest challenges on the way to starting a business?
There is no challenge in finding an idea. There are thousands of problems waiting to be solved, and each of these approaches has several teams working on that one idea. The real crux is building a human relationship with whom you can go through thick and thin. For me, it was challenging to find a co-founder with complementary skills to tackle all the challenges.
Here, Entrepreneur First did a great job building a platform for talents that gave Ceyhun and me the chance to meet. We had already hit it off before we officially started and then quickly realised that we both complemented each other so well that our productivity formula meant “1 + 1 = 3”.
After we had joined forces, the German love of bureaucracy and digital ignorance was indeed our worst adversary. Things have changed, especially in the financial sector. Who is sending contracts across the continent by post? That’s a thing of the past! It is great to see that EBS Alumna Jessica Holzbach has also built a solution with Penta. Everything is doable with like-minded partners.
On 16 September, EF has organised a large-scale Demo Day. Would you please tell us more about that?
The Demo Day is the major public presentation of the start-ups from the Entrepreneur First (EF) portfolio. It is also their first appearance before the approximately 2,000 VC investors who belong to the network. The start-ups present their case, similar to what we are used to from big names like YCombinator. There will be a different EF website for this, where you can find information after the Demo Day and contact us.
EF is an established talent investor whose investors include LinkedIn founder Reid Hoffmann and Peter Thiel’s Founder Fund. For almost ten years now, they have made it possible for individuals to find a co-founder with similar interests to join forces to build up a business. The value of the portfolio is meanwhile at over 3 billion US dollars.