If the year is 2025, why are we still developing like it’s 2020?

In AI Insights for Tech Leaders, we bring together key takeaways from Microsoft Build and Google I/O, along with real-world use cases from GX2, showcasing Artificial Intelligence in practice.

In May 2025, during the Google I/O and Microsoft Build events, one thing became clear: AI is the new foundation of technology, shaping the way we interact with and build systems.

Our CEO, Luiz Garcia, has been discussing this “AI-First” future with clients such as Unimed do Brasil. In this edition, we highlight the most relevant developments in the AI landscape.

Gemini 2.5 and Deep Think: Google’s New Era of Personal and Proactive AI

Google is entering a whole new dimension with Gemini 2.5 and the Deep Think Mode. Gemini is no longer just an assistant; it’s a partner that learns from you, provides personalized solutions, and optimizes your daily workflow in a personal and proactive way.

What is Deep Think Mode?

Deep Think Mode allows the model to perform deep and complex reasoning before generating a response, thinking “out loud” and exploring multiple reasoning strategies, much like a human expert solving a challenging problem.

How it works:

• Considers multiple hypotheses before providing an answer

• Explores parallel strategies to optimize results

• Adjusts reasoning intensity based on task complexity

• Improves accuracy and quality in challenging tasks

Use cases:

• Advanced mathematics

• Software development

• Deep research and data analysis

• Scientific reasoning

Deep Think Mode makes Gemini 2.5 a powerful ally for companies seeking strategically applied AI.

AI Agents and Turbocharged Copilot in Microsoft 365

Microsoft is revolutionizing productivity in Microsoft 365 with AI agents and a turbocharged Copilot, which can:

• Work autonomously

• Optimize tasks and suggest solutions

• Personalize experiences based on your company’s data

Jules: Google’s AI Coding Agent

Jules is an asynchronous AI coding agent, released in public beta in May 2025, capable of performing software development tasks independently, directly integrating with GitHub repositories.

GitHub Copilot as a “Peer Programmer”

The GitHub Copilot now acts as a true “Peer Programmer,” working side by side with developers:

• Automatically fixes bugs

• Suggests intelligent solutions

• Frees the team to focus on innovation and creation

Challenges for AI Adoption

According to IDC, cost and data control are still the main barriers to AI adoption.

Starting with POCs (Proofs of Concept) is the safest way to:

• Validate ideas before heavy investment

• Measure the real value of AI for the business

• Adjust approaches with agility

• Avoid wasting budget and time

This type of experimentation is exactly what the GX2 Innovation Lab promotes.

GX2 Innovation Lab: Experimentation and AI POCs

At GX2 Innovation Lab, we co-create POCs that help companies prove value before scaling, combining:

• Expert technical support

• Strategic business vision

• Cutting-edge tools for AI, Low-Code, APIs, and Data Platforms

GX2 POC Examples

1. Smart Training and Onboarding
A U.S.-based company developed a solution using generative AI and smart search, capable of:

• Locating specific parts of training videos

• Answering questions and generating strategic insights

• Supporting the creation of documentation such as ISO 9001-aligned SOPs

2. Optimized Bid Analysis
A major Brazilian corporate benefits company created an intelligent bot using OpenAI and Google Gemini, acting as a Bid Analyst, reducing analysis of large volumes of bids from days to minutes.

Want to test something like this in your company? Schedule a conversation with our team.

Conclusion

The future of technology is AI-First, and the time to experiment with AI is now. Start with POCs, test ideas on a small scale, and discover how Artificial Intelligence can transform productivity, innovation, and decision-making in your business.

Talk to our specialists and learn how to strategically implement AI in your organization.