IT IS A HARD IT'S LIFE

The advent of AI for businesses has an undeniable impact, especially on business and technology areas.

Before the pandemic, it was estimated that there would be a shortage of 240,000 IT professionals in Brazil alone in the following years. Post-pandemic, with downsizing and crisis in many sectors, leading to mass layoffs, this number seemed to be under control.

Then came 2024, and ChatGPT unveiled a new reality. From the world without programmers, envisioned initially, to the current reality, much has changed. What hasn't changed? The mismatch between structures.

Let me explain. Before AI, most companies suffered from a chronic problem: technology departments that couldn't deliver solutions in the timing and format desired by the business areas. This problem, which I addressed in a previous text, had three sources:

• Lack of qualified IT professionals, as mentioned above.

• High cost of professionals, mainly due to scarcity, which led to companies rationing their workforce.

• Total lack of alignment between IT and Business, creating constant tension between the areas. Business wants time to market, and IT wants bulletproof solutions.

I'm not including the overvaluation of bad IT professionals in this equation, because that's a problem we have not only in this segment. In Brazil, there's a shortage of good professionals in almost all areas.

Things started to change in companies before AI, with low-code and no-code tools, which allowed professionals from various areas, with a little dedication and goodwill, to build some simple and useful solutions.

The biggest problem of the lack of alignment between IT and Business has already begun to become apparent and, in my opinion, will only worsen in the coming years. The reason? While someone was developing a tool to manage emails using Bubble or Flutter Flow, IT continued to navigate the turbulent waters of legacy systems, with complex databases and code, believing it was in a separate world.

With AI, more specifically with complete AI agents that integrate interface, application, database, and server, the business has changed dramatically. Suddenly, business solutions are no longer peripheral and restricted: they begin to interact with the entire ecosystem of companies.

It's no use having a robust legacy system with complex databases and a huge IT team if none of that communicates with these solutions that will multiply. For years, companies have sought to build secure, complex, watertight, closed legacy systems with protected and isolated databases. None of that will survive in the coming years, and the fault lies not with AI, but with the architecture: these systems are desktop versions of mainframes of the past. Due to security-faulty software like Windows Server, companies have built exoskeletons made of restrictive and limiting applications. As we learned in biology, in order to grow, animals with exoskeletons need to shed their shells and become vulnerable until a new armor solidifies. That's what will happen to companies, if they don't die first.

I'm proposing something more practical and intelligent: abandoning these armored barriers in favor of APIs and integration components that allow the technology ecosystem of companies to overflow into all business areas, both existing and future, integrating with fast, practical solutions dedicated to business needs that arise over time.

It is up to IT to build both the interoperability that already exists in giants like Google and the policies and barriers that will prevent data theft, viruses from being implanted, and servers from being hacked. IT departments must implement these exchange environments to allow, for example, the technical area of ​​an insurance company to build a risk-focused BI for its auto insurance portfolio without any proprietary or confidential data being accessed or stolen. To allow marketing to send WhatsApp messages integrated with app notifications to 32 customer segments to register for a new service without any sensitive data being compromised.

For this to happen, hard-code programmers will have to either submit to LLMs or coexist with programmers native to this new reality. I see new developers still resisting introducing AI into their daily work, as if they were opening a door for an enemy to enter. AI is already inside companies at all levels, from the image that communication requests for GPT Chat to presentations made to a client by the sales department.

Delaying this process will spell the death of many businesses. Tools are tools, and AI is a super ultra tool that will reinvent many things. The saying goes that it's better to keep your friends close and your enemies closer.

IT needs to stop looking at code and prompts and start looking outside its castles of servers and cables and look at what, for years, has been the villain of all the systems already developed: the user.