AI Companies, but not for long

Nearly all companies claiming to be AI companies today, including mine, would not call themselves AI companies in a few years. There's a fascinating transformation underway, and understanding its nuances not only reveals the trajectory of AI but also unravels key insights into the dynamics of startups and technology.

Which companies will still identify as AI companies? The B2B startups and companies selling their AI models through APIs or offering the infrastructure to run AI models to other companies will. The rest won't. And here are the reasons. (If you decide to leave, read the last paragraph before doing so.)

Eventual Ubiquity

Picture this – someone proudly proclaiming they have a 'website' for their business. Does that raise eyebrows or spark excitement? Not really. Having a website has become a baseline expectation, much like how businesses, big or small, incorporate software in various facets. Not every such entity is an internet business or a tech company. Even when a small mom-and-pop store owner claims their business listing on Google Maps, they are indeed using the Internet to augment their business and increase foot traffic. They might approve the listing over a call without using the Internet and don't own the technology, but they are using it nevertheless. They might not be paying in cash for the listing, but they are paying Google by making it more accurate and useful. Or consider a solopreneur monetizing their decent social media following – that social media profile is their use of the software! The software doesn't have to be paid for or built in-house. And neither you nor those store owners or influencers or every business with a website think of themselves as an internet company. Similarly, a decade from now, the widespread use of AI in products and even solo ventures will be as pervasive as having a website today. Still, those companies won't be considered AI-centric. I've chosen some extreme parallels here deliberately to get the point across.

Some companies use the Internet more, while others use it less. A local store owner who takes orders and gives updates over WhatsApp is using it less. The other local store that chooses to use online presence, inventory management, online payments, loyalty programs, and cloud-based CRM is using it more. Those who use it more will have more success. But both are using it. All tech companies are certainly using it in some capacity. Similarly, AI-based health tech or AI-based ERP in the future will simply become health tech or ERP companies, whereas those who don't use AI at all will perish. When a technology or a tool becomes the backbone of every other tool out there, it ceases to be mentionable, and mentioning it becomes absurd. Marc Andreessen said in 2011, 'Software is eating up the world.' Perhaps AI is pulling up a chair to the table. Just as we no longer categorize businesses by their internet or software usage, AI will fade into the background, becoming an assumed component rather than a standout feature. Software and its very important subcategory of AI have now started dining together.

When I fill out the LinkedIn page or any other form for my company, the only suitable category I find for it in most places is 'Internet company,' which feels weird. I'm sure I'm not alone in this. There should be categories by now. I can pivot through 50% of the available and feasible startup ideas today and would still have to select an 'Internet company.' Mentioning that tells the reader nothing about my company. A decade from now, telling someone you are an AI company using the current usage patterns will tell them nothing about you.

Entrepreneurs often get called out by quality investors for calling themselves a technology company when they are not. You might have seen this happen on Startup-based reality TV shows, but it happens with a much higher frequency behind the scenes. A decade from now, something similar will also happen to the companies calling themselves AI companies. But until that shift occurs, embracing the AI label remains crucial. More on that in the final paragraph.

Another important factor is that AI has a wide range of applications. It will touch and enhance every software and sector out there, including search engines, social media, healthcare technology, robotics, research and development, etc. In that way, it is much like the Internet, which has again improved everything. A more fundamental example is electricity. Electricity didn't just impact a single sector. Imagine calling yourself an 'electricity company' as a manufacturing company in the early 1800s. Sounds absurd now, doesn't it?
Note that I am not downplaying AI. In fact, I am doing quite the opposite. I am saying that AI will be as ubiquitous as the Internet or electricity. Thus, the insight that it will be ubiquitous is important for companies to incorporate it wherever suitable.

Customers Don't Care

Another way to think about this is that for most 'AI' companies, AI just sits in the backend, making everything seamless. AI is a means to solve a problem. The same problem was solved earlier through some other class of algorithms, even if inefficiently and not as well. Despite AI's superior problem-solving capabilities, most customers, except for a handful of early adopters, are indifferent to the backend mechanics. The customers don't care whether you are using AI, whether your perfectly written code runs on the best-in-class server architecture to return what they want to see, or whether you have an exceptionally well-trained and humanely treated monkey in the server room magically returning the correct answers. They don't care which language, server, framework, or tools you use to find the correct answer. They only care about the experience. For all they care, AI is just one such monkey.

The insight that most customers don't care how your product is working behind the scenes is vital to understand because entrepreneurs, especially those with a heavy tech background, can often fall into the trap of focusing on how their product is working and what it is built with rather than how or even whether it solves the customer's problem. Unless that behind-the-scenes tool gives you a 10x advantage in some way, picking the simplest, easiest, and quickest solution is more important than the latest one.

AI undoubtedly gives most companies that 10x advantage for now, justifying its use in the startup name... for now. But a decade from now, calling yourself an AI Tech startup would be akin to identifying as an Internet Tech Startup in the 2010s or a Mobile-first Startup in the 2020s, which was fine to do a decade before it wasn’t. Until then, using the term AI startup will help customers, investors, media, and the general public better understand the startup. It helps you immediately convey what the USP or the secret sauce is. It immediately tells them you are trying to build something cutting-edge, saving everyone's time and increasing response rates or conversions. So it's okay to use it… until it won't be, so know the caveats :).