ChatGPT and many other tools that have been appearing based on Large Language Models, these large linguistic models of Artificial Intelligence, are just that: models capable of understanding and making themselves understood in natural language. Given that a significant part of what companies do in their daily activity is some form of written or spoken language, Artificial Intelligence has the potential to become involved in most of our tasks.
The pace at which this will happen, we do not know. According to the 2021 Digital Intensity Index (DII), only 56% of European companies are at the minimum level of digital intensity, which includes the use of at least four out of 12 digital technologies (such as Artificial Intelligence). Similarly, when it comes to the adoption of cloud computing , only 4 out of 10 companies in the EU were using this technology in 2021 (although this technology has been available on the market for several years).
Of course, business reality is often prone to lengthy technology adoption processes. If SMEs often struggle to gather the resources needed to adopt new technologies, large enterprises are built on complex infrastructures, composed of diverse platforms and applications, which make it challenging to integrate new solutions. While it is not correct to assume that past observations will be repeated in similar contexts in the future, it is reasonable to contemplate that these Artificial Intelligence tools may still take some time to be fully adopted in the various areas of the day-to-day business.
If so, the realization of the most apocalyptic perspectives regarding the impact of AI on the labor market is not around the corner. Instead, what we already see happening is time savings of around 40% and quality increases of around 18% in ChatGPT-assisted tasks.
Companies should therefore start prototyping and experimenting with the use of these tools now. They can do this by conducting small trials with controlled use cases.
A good way to start this process will be, for example, through internal incremental solutions: the creation of a chatbot to provide information for sporadic consultation - for example, the invoicing process to suppliers, the data of the company's financial vehicles, etc. more specialized tools to support market research, which enables the creation of personas that represent our target audience. Reviewing business plans, writing SEO-optimized articles, or even getting more specific insights into finance. There is no shortage of tools and guides to make the most of them.
These incremental solutions, from existing tools, circumscribed to specific employees or teams, will allow us to understand which use cases have real value for the business. In parallel, companies should think about what disruptive solutions can be created with this technology. These two parallel paths feed into each other, helping to create a culture of experimentation and joint exploration.
There are, however, two aspects that must be observed: these initiatives must be promoted with the supervision of IT, so as not to create small niches of shadow IT, which can have consequences for the entire application in the organization. And, above all, employees must be trained and sensitized to the use of these tools, so that they can discern the type of information appropriate to work on these platforms. After all, according to a recent study by security company LayerX, more than 15% of employees post internal information on ChatGPT, of which more than 20% is confidential or sensitive.
While access to these tools is easier than ever, it's also very easy to get sidetracked: according to a Pew Research study published in May this year, 58% of Americans say they are familiar with what ChatGPT is. However, only 14% admit to having tried it.
Companies need to make a deliberate investment in experimenting with these technologies. Without experimentation, it will be difficult to understand the true potential or implications they present. If they do not seize this potential for competitiveness, some competitor will.
