OpenAI launched Chat GPT at the end of 2022 and the world would never be the same. The witty and useful interactions seemed to laugh in the face of any Turing Test, and people took a closer look at technology like Large Language Models (LLMs) and Generative Pre-training Transformers (GPTs). AI companies skyrocketed in value, keynotes became all about AI, and everyone was left wondering: What’s next? In this week's newsletter, we cut through the noise and explore the real innovations that have actually materialized since then.

To do this, let's first examine what we mean by real innovation.

🎯 A Product That Serves a Need – No Hype!

In AI, every day brings news of large investments, new features, strategic cooperations, senior staff changes, product upgrades, and share-price highs. Often, these come with a label of innovation. But are they really? In the flood of announcements and superlatives, we use two criteria from design thinking to separate hype from real innovations:

1. There must be a product. One that people use and ideally, pay for. Amidst the hype, people are often drawn to investments and visionary products that are still potentially years away. In this analysis, let us thus see what is already available today. We deprioritize headlines of investments into AI companies or products still far away from being launched.
2. The product must serve a real need. We are looking for AI solutions that solve real problems and add genuine value. Simply adding AI to your business to ride the hype wave isn't enough. Also, we are not interested in platform wars. i.e., it does not matter whether you use Zoom, Teams, or the next AI-powered video conferencing app if you still use it to do the same thing. We're interested in innovations that address actual pain points and improve lives in meaningful ways.

As a side note, this doesn’t mean that investments in AI companies are poor choices or that platform wars are irrelevant. On the contrary, fortunes will be made in these areas. Just look at NVIDIA 's continued meteoric rise to becoming a USD 2.5 trillion (!) company selling chips that AI learns and runs on. But today, we focus on real innovation in the sense that AI is the product, it made it to market and people are using it today.

4 Examples of Real AI Innovation Today

Let's dive right in and explore the top four AI innovations that have emerged since the launch of Chat GPT sparked the current wave of AI enthusiasm.

1. Chat GPT ✍️

OpenAI's Chat GPT, with around 180.5 million monthly users as of May 2024, is the most prominent and visible AI product in recent times. It's the one use case people know, have on their phones, and interact with daily. It exemplifies real AI innovation by providing practical utility and transforming how we interact with technology.

2. Adobe Firefly 🎨

Adobe Firefly is a suite of AI-powered creative tools that integrate various AI technologies, including LLMs, to enhance and streamline design processes within Adobe’s software ecosystem. The most important features are text-to-image generation and generative fill. It's a great example of how AI can be deeply integrated into an existing offering, very different from the standalone Chat GPT. While this makes it harder to cut through the hype and evaluate its standalone capabilities and revenues, its success within the community is considerable.

3. Klarna Customer Support 📞

Klarna , the Swedish fintech company, is an early adopter of AI and a great communicator about how its journey is developing. CEO Sebastian Siemiatkowski himself is very vocal about integrating AI into all aspects of the business. The Klarna AI assistant for clients, powered by OpenAI, is a prime example of a practical AI application. At the beginning of 2024, it famously handled two-thirds of customer service chats in its first month, doing the work of roughly 700 full-time agents. Additionally, Klarna provides its employees with a GPT-powered AI assistant called Kiki, which successfully handles around 2'000 inquiries per day.

4. AlphaFold 🧬

AlphaFold, developed by Google DeepMind , has revolutionized protein folding predictions. How proteins fold is a crucial and up until now very expensive area of research. There is a light version of the tool and database available online, which is said to have 1 million users in 190 countries. Additionally, Alphabet Inc. ( Google 's mother company) is commercializing the technology through Isomorphic Labs , which recently announced partnerships with Eli Lilly and Company and Novartis worth up to USD 3bn. The UK company recently opened a second office in Lausanne. AlphaFold and Isomorphic stand out as a real AI innovation by solving critical scientific challenges and enabling advancements in medicine.

Conclusion

These examples illustrate the transformative potential of AI in our daily lives as well as in the B2B space. They show that the AI revolution is already here, and happening in various forms and speeds all around us. We were quite strict with the "show me the money"-style of evaluating which AI innovations are real. Consider these 3 reasons why the list above turned out quite short with only the top cases making it:

1. Rapid Evolution. As you read this, both minor and major revolutions are being planned and rolled out. But as they are not out now, they did not make the list. For instance, Apple 's Siri gaining GPT-level capabilities and accessing all our apps might be a groundbreaking innovation just around the corner.
2. Broad Spectrum. While the current excitement centers around large language models (LLMs), many innovations have been leveraging AI for years. Think of how Tesla has been developing computer vision and autonomous driving. LLMs and Chat GPT have acted like a rising tide, lifting all boats in the AI landscape. We prioritized new LLM/GPT cases.
3. Market Dynamics. AI companies are flush with cash, and new opportunities continue to emerge. OpenAI itself was famously not trying very hard to monetize its product at the beginning. And still today, companies are largely fighting for user adoption and market dominance while promising a great product that users want to buy further down the road. We focused on the ones that have paying customers today.

One and a half years after the GPT shock, it may seem that the market is still dominated by visionary announcements and large investments, with relatively little to show for today. While our four examples today demonstrate that real innovation with AI is possible, the future will determine which initiatives manage to create actual products that people want and which will pop like a soap bubble.

What do you think? Did we miss any of your favorite AI tools? Share your thoughts!