It’s a genuine question what AI's killer app/platform/solution is - we do not know the answer. But we bet you don’t know it either. At best, you have guesses. But we can certainly say that this uncertainty is universal and it has important implications for SMEs which we’ll tell you why in just a couple of paragraphs.
Anyone remember what it was like to develop an app for iOS before the App Store? Apple's original strategy was to not allow third-party apps to be installed on the iPhone to ensure the integrity, security and quality of the product and the apps that ran on it. As is often the case with ill-considered bans, here too the rules started to be bypassed, with 'jailbreaking' soon to start. Jailbreaking involved circumventing Apple's software restrictions on iOS devices, granting users root access and the capability to install unofficial third-party apps. Users had to install applications from unauthorized sources, voiding the warranty of the device and creating potential security risks. This solution wasn’t exactly bright for developers either. They had to host the app, find a place to sell it, and create a step-by-step guide for their customers to help with the installation (you can imagine how much developers loathe these kinds of tasks). Apple consistently tried to patch vulnerabilities that facilitated this practice, but the new era of telecommunications started when they realized they had to meet this market demand, not fight it.
Apple released their Software Development Kit (SDK), which has made it easier for developers to create apps, and has also ensured the quality standards that were originally intended to be protected. The App Store was introduced with only 500 apps in 2008. Today there are 1.8 million apps in the App Store and 3.7 million in Google Play. It is safe to say that this has become the dominant way to distribute software.
Back to AI. Today, in this respect, we may be in the post-iPhone pre-SDK era. If an SME today decides to develop an AI-based product, it will face very similar difficulties as a coder developing an app for iPhone in 2007. The difference is that developing an AI solution is much more expensive. (For more on how great an idea it is to develop your own LLM, read our previous blog post here.)
For the time being, the killer apps/platforms/solutions in the use of AI are mostly at the end-user level. These solutions cannot be ignored, they are becoming real factors in content production, whether we are talking about images, video or text: ChatGPT, Midjourney, DALL-E, etc. There is also a second line of AI solutions for SMEs, including some CRM, marketing and advertising, finance, HR tools, and virtual assistants. All of these can of course be used to do corporate work. But do they provide a solution to the question: what would it take to get AI solutions into SMEs' products? (See more at eDiscovery Today and Medium.)
We believe this one important link is missing in the current AI market. Some use cases exist, but there is no platform yet to make AI-based product development accessible to a wider audience. End users can already choose from a set of ready-made functions, but this is not the same as an AI package available to SMEs for developing their products. (Although Open AI launched its ‘app store’ this year, only time will tell whether this is the real deal or something else.)
And yes, we think that the SME sector is essential for a well-functioning 'AI market': they are close to their customers, have a lot of innovation potential, and their different operations from large enterprises allow them to deliver solutions in niche areas. What should SMEs do until this missing link is clearly identified? Read our latest white paper on the topic.
Innospective (2018): Why the App Store was more important than the iPhone
The Verge (2013): How Apple changed the world again: the App Store turns five
Open AI (2024): Introducing the GPT Store
Comments