How 99x is building Agentic AI technology from Portugal (interview at Teamlyzer)

Interview with Guilherme Ramos Pereira, Head of Strategic Development at 99x Portugal (original piece in Portuguese published at Teamlyzer)

 

From consultancy to product: 99x is looking for developers and engineers in Portugal to build xians.ai, an AI agent orchestration platform.

Cleverti began 15 years ago as a nearshore consultancy in Lisbon. In 2023, it became part of 99x, a global group with over 600 professionals, founded in 2004, and continuously expanding its range of services, products, and technological capabilities. That path naturally evolved into building proprietary technology. The result is xians.ai, an orchestration platform for AI agents already being used in production across multiple sectors.

We sat down with the 99x team in Portugal to understand how a platform like this is built, what technical skills they are looking for, and what opportunities they see for the Portuguese tech community.

To better understand the current reality of 99x, we spoke with Guilherme Ramos Pereira, Head of Strategic Development.

 

1. For those unfamiliar with agentic AI: what does xians.ai do that’s different from using ChatGPT or Claude directly?

When you use ChatGPT or Claude, you’re interacting with a single model. It’s excellent for many purposes, but it has limitations when you need to automate complex processes that involve multiple steps and multiple systems.

xians.ai works as an orchestrator. It coordinates several specialized agents working together to solve complex business problems. One agent monitors information sources. Another analyzes data. Another interacts with internal systems. All of them collaborate autonomously to complete full workflows.

That’s the difference: you’re not using AI just to answer questions, you’re using it to execute end-to-end processes that previously required constant human involvement.

 

2. The integration of Cleverti into the 99x group in 2023 enabled a transition from traditional consultancy to building product. How did that shift happen?

It was a natural step. For 13 years, we built a strong foundation in nearshore, with a deep focus on technical quality and close client collaboration. When we joined 99x, we kept that identity but gained the scale and capability to go further.

Being part of a global group with more than 600 professionals gave us access to expertise and resources to build our own technology. The transition was not about abandoning consultancy, it was about evolution. We still work with clients every day, but now we also create complete solutions we can scale globally.

 

3. You say xians.ai works as an orchestrator of cooperating agents. How does that work in practice?

Imagine a complex business process that would normally require several people working in sequence. With xians.ai, we design specialized agents, each with specific responsibilities, and the platform coordinates how they collaborate.

One agent collects and monitors external information. Another analyzes and makes decisions based on pre-defined criteria. A third interacts with internal systems or executes actions. The platform ensures they work together, share context, and move toward the final goal.

The power is in autonomy: agents adapt to context, learn from what they encounter, and make decisions intelligently. It’s the orchestration of intelligence, not just tasks.

 

4. What types of problems are companies looking to solve with autonomous AI agents?

Companies want to automate processes that consume a lot of time but still require decision-making and adaptability. These are not repetitive tasks, they involve analysis, interpretation, and coordination between systems.

The pattern is clear: free teams from heavy operational work, reduce costs without reducing quality, and scale operations without scaling teams at the same rate.

Ultimately, companies want solutions that combine efficiency with intelligence, while maintaining control and security over critical processes.

 

5. How do you ensure that autonomous agents running continuously remain reliable and don’t make wrong decisions?

Reliability starts with design. We don’t let agents operate freely without boundaries. We define clear action limits, structured decision criteria, and validation mechanisms throughout the process.

Our approach combines AI with human oversight at critical points. Agents have autonomy to execute tasks, but important decisions can be reviewed, and there is transparency about how and why each action was taken.

We always operate with this principle: these systems serve people, they don’t replace them.

Security, scalability, and trust are the pillars of everything we build.

 

6. You mention teams made up of developers, engineers, analysts and AI specialists. What kind of technical profiles are you looking for in Portugal?

We look for people who want to build meaningful projects. Technical competence is essential, but what really matters is curiosity to solve complex problems and the willingness to work with emerging technology.

We need developers, engineers, and analysts who don’t just follow specifications, but who think about architecture, scalability, and real business impact.

For AI specialists, we want people who understand not only the models but how to apply them to real-world business problems.

Portugal has strong technical talent. What drives us is creating opportunities for that talent to build real product, cutting-edge technology, without leaving the country.

 

7. For someone coming from traditional consulting: what changes when you start building an AI product?

The biggest difference is mindset. In consulting, you solve a problem for a client, deliver, and move on to the next project. In product, you build something that will evolve continuously, scale, and be used by multiple clients in different contexts.

Technical decisions you make today have long-term consequences. You must think about architecture, maintainability, and scalability much more rigorously.

And in AI product development, the field evolves extremely fast. Continuous learning is essential.

But the most important shift is this: instead of delivering one project, you’re building capability. You’re creating technology that solves problems for many companies, not just one.

 

8. What technical skills are essential today for anyone who wants to work seriously with agentic AI?

First, a solid foundation in software engineering. Agentic AI is not just prompt engineering, it’s building complex, robust, scalable systems. You need architecture, APIs, state management, error handling.

Second, a real understanding of how language models work: limitations, context, and how to structure inputs for consistent outputs.

Third, experience with automation and orchestration. Agents need to interact with existing systems, databases, external APIs.

And lastly, something less technical but absolutely critical: the ability to translate business problems into technical solutions. Knowing where AI adds value and where it doesn’t.

 

9. Do you think Portugal has the conditions for more tech companies to move from service delivery to developing their own products?

Yes. Portugal has strong technical talent. That was never the issue. What changed was the ecosystem: it matured, gained experience, and built meaningful references.

We spent decades delivering services to international markets, and that built real competence. Our teams work with demanding clients, solve complex problems, and master advanced technologies. That experience is the foundation needed to build product.

The conditions exist: talent, technical maturity, and access to global markets.

What’s often missing is ambition, capital, and the ability to scale.

Our experience shows it’s possible. 99x is building proprietary technology and selling it globally. And Portugal should be recognized not only for the talent it exports but for the technology it creates.

 

10. Now that you’re on Teamlyzer, what kind of contribution do you want to bring to the Portuguese tech community?

We want to share what we’re learning. Agentic AI is a new area. Lots of hype, little real production experience. We have hands-on experience building and operating these systems, and Teamlyzer is the right place to share it with the community working on similar challenges.

We also want to create real opportunities. We’re growing in Portugal and hiring. Teamlyzer connects us directly with the talent we’re looking for. People who want to work on AI product development without needing to emigrate.

And finally, we want to show that it is possible. Portugal has top-level technical competence, but historically we’ve exported more services than product.

The best way to change that isn’t speeches, it’s building technology that works and sells globally. Being on Teamlyzer allows us to showcase our team, our culture, and to stay close to the largest tech community in Portugal.