AI adoption in businesses: Breaking the glass ceiling

All companies today have realized the power of Artificial Intelligence (AI). Professionals across domains are talking about how they can leverage AI to boost their business. AI has been instrumental in transforming online advertising, online product suggestions, trend predictions, customer-facing automation and more for several organizations. Many companies have integrated AI into their businesses and more companies are working tirelessly to extract the potential value AI can bring to them. While everyone in the world is busy adding their colours to the canvas of AI use cases, in some corner of this canvas, there lies an empty space where several businesses thrive that are still untouched by this revolution.

One must ask, when the biggest technology companies are developing state-of-the-art AI systems that are, to an extent, available for free, why don’t we see businesses around using AI? Why don’t Kirana stores use AI when it could possibly help them optimize their inventories? Why don’t sweet shops use AI when it could help them optimize their production? Why don’t small-scale manufacturers use AI when it could help them optimize their supply chain? The answer to all these questions is not unawareness, resistance to change, or even lack of confidence in AI, it is the high costs of implementing such systems. Today, if a company wishes to implement an AI solution to address a problem, it is highly likely that they will have to first develop it. This will require allocating funds for hiring a group of developers, gathering relevant resources, and dedicating enough time for the project to bear fruit. This is precisely where several organizations take a backseat in their efforts to embrace Artificial Intelligence. The problem is largely financial in nature. The question most businesses struggle with is “How will the costs of implementing such systems be justified?” rather than “Whether to implement AI systems or not?”. This is the “glass ceiling”. When seen from the outsider’s lens, this idea seems a fallacy as the technology is largely free with products like ChatGPT and Midjourney readily already available. This is where the catch lies. Only a “few” such products are available in the mass market. There still exists a huge gap between the ones who are developing this technology and the ones who should be using it.

The logical question that follows is “How do we break this glass ceiling?”. Given the complexity of the technology, we need more middlemen who would take the technology to the masses. We need more and more products that bridge the gap between the “developers” of AI and the “users” of AI. The task is to make Artificial Intelligence accessible to people in a manner that is comprehensible.

This can be achieved only when we build an entire AI ecosystem, accessible to all. Big tech companies need to develop “platforms” that can form the basis of the development of other products. Such platforms should allow the user to comprehend the technology while at the same time providing enough resources to build something more. “ChatGPT” is an excellent example of such a platform, as one can easily explore the applications of Large Language Models and access resources in the form of APIs and developer documentation. These resources can be used to build a “Poe.com”, a product that allows you to build your own chatbot using ChatGPT. After platforms, the next required set of products is the “standard products”, products that solve a standard problem common across companies and have wide applicability. These standard products may or may not form the basis of the development of other products. Video editing is an example of a task that contains multiple standard problems prevailing across companies in the domain of content creation.

“Invideo.io” is a company that develops multiple AI solutions to address many of these problems like script generation, voice modifications, and snippet generation. All these products are good examples of standard products. Finally, arrive the “specific products”. These products are intended to solve a particular problem being faced by a specific industry or company. Needless to say, they exhibit a high degree of customization. These products are few and are scattered across industries as they may not be widely popular. Often these are custom-built and leave no scope for further development, i.e. one cannot use these products as the underlying architecture to build another product. Issues faced by the insurance industry and banking industry can be examples of specific problems that can be solved by such products due to the severe regulations in these industries. Problems faced by SMEs and MSMEs often fall in the same bucket as they generally tend to cater to a particular category of customers, thereby addressing a particular category of problems. Even the problems faced by local Kirana stores are quite specific to be solved. To address all such problems, one needs to create customized solutions using other products or raw technology as their underlying architecture. Such products may employ low-code or no-code development methodologies to ease out customization.

It is to be noted that all three categories of products required to build this ecosystem need not necessarily be interdependent on one another, though, in practice, a level of interdependence might be observed. The figure below explains the same. Platforms are directly dependent on the raw technology whereas standard products are directly dependent on platforms and indirectly on the raw technology. The strength of dependence varies with different categories of products. The existence of all products together will create the required AI ecosystem. The world is moving in an unprecedented manner and businesses are finding it difficult to keep up with the pace. The use of the right AI products can offer improved efficiency and hassle-free operations. Though recent developments promise a bright future for technology, it is yet to mature. The only way to improve the adoption of AI is by accelerating the development of products across categories.