*By Frederick Heitmann
Over the past 40 years, we have witnessed the emergence and rapid evolution of the Information Technology (IT) department, driven primarily by the rise of the internet. Now, we are seeing a similar movement with Artificial Intelligence (AI). Companies that do not yet have a dedicated AI department, either in-house or outsourced, are falling behind in this ever-changing landscape.
It is important to highlight that AI is not a buzzword recent. Since the emergence of computing, the concept of AI has been discussed, since the first projects. AI has been a reality for a long time.
To discuss generative Artificial Intelligence, we need to understand the concept of buzzword (buzz – noise, word – word), which is used to describe a new technology that emerges, as we have already seen with blockchain, metaverse and social networks, among others. In 2023, we were impacted by a new buzzword, Generative AI. However, are we really facing yet another fad?
In the current market, there are already several AI applications, including generative ones. IBM Watson, for example, was experimented with even before the emergence of GPT. While not generative per se, it does have an important aspect, which is context. With the success of the GPT, OpenAI has opened the door to extensive research in colleges and companies, enabling rapid adoption of this technology.
Major market players are developing their own AI solutions. We are at an inflection point and maturity with regard to AI, in many forms, including the GPT. However, technological progress has outpaced business maturity, which has only recently been impacted. People are getting interested, experimenting and doing proof of concept. The market is gaining knowledge about the application of AI and developing this knowledge to apply it effectively.
However, I believe that an educational process is still lacking to prevent this new technology from being absorbed just as a passing fad. Gathering experts and technical professionals is the most assertive approach to separating the wheat from the chaff in business. Therefore, it is crucial to have a department dedicated to AI, whether internal or outsourced. Generative AI cannot simply be turned on an instant messaging platform and used immediately. Response time is not adequate as it takes time to generate a relevant response.
It is a fact that AI models are not yet fully developed. For example, creating an e-commerce page requires a response in milliseconds, while Generative AI can take a few seconds to generate content. Therefore, an adaptation process is necessary to identify what is really important in terms of usability, requirements and real application for the business.
When we talk about Generative Textual AI, we must understand that texts contain less information than images, which implies a greater volume of information to obtain relevant results. There is much to be discussed and understood in this regard.
There are already several alternatives to GPT on the market, which allow you to chat with different personalities, such as movie characters, celebrities and politicians. In addition, there are solutions capable of creating digital content, such as ads for social networks, scripts for videos and email marketing. These products have the potential to revolutionize the way products are offered on e-commerce platforms and how consumer interaction takes place. However, it is crucial to deeply understand the real value of these applications to the business.
It is not yet clear to the market what is the level of maturity of each possibility of Artificial Intelligence applied to these tools. The AI that deals with images, for example, is more advanced and can be easily trained to perform augmented reality in architecture and decoration projects, measuring spaces and identifying features. However, in other applications, there is still a long way to go.
First of all, we need to understand how to train the GPT to assimilate specific information about a company and its products, ensuring that it is not creative or autonomous in its responses. Contextual language needs to be human and not creative in relation to business expectations. If we eliminate AI's creativity, it ceases to be human; on the other hand, if we increase creativity, it may give incorrect information about the company. The big current challenge is to train this AI so that it embodies the brand's personality.
With regard to omnichannel, AI can be an ally in solving this challenge as e-commerce evolves. We already have AI capable of building storefronts, improving search results and creating ads. However, we still face a significant problem: the integration of all the artificial intelligences involved in the e-commerce process in a single context.
We've been looking for the real one for a long time. omnichannel, and Generative AI can help us achieve it. When a consumer buys a product, he creates a relationship with the brand, and AI needs to understand and respond to the entire purchase journey, ensuring a complete experience. Currently, AI assistants are trained to deal only with superficial questions, they have not yet reached this level of sophistication. My vision for the near future is for AI to become the central contextualization tool, integrating brand, product and customer across all channels.
The GPT has the potential to act as a link that helps build this entire scenario. Until now, the pieces were disconnected. You can have an AI to build windows, but it does not automatically connect with the Artificial Intelligence responsible for advertising, for example, or customer preferences.
The message for e-commerce companies eyeing Generative AI is clear: it is urgent to deploy a department dedicated to Artificial Intelligence to acquire knowledge. Investing in science, research and development is fundamental. If you're not studying the topic, you're terribly late. Competition in this area is accelerating, and it is necessary to keep pace, investment in knowledge is essential for success in this new era of technology. It became a race, believe me!
*Frederico Heitmann is Innovation Director at Quality Digital
Notice: The opinion presented in this article is the responsibility of its author and not of ABES - Brazilian Association of Software Companies
EN 
PT 
