Inside Hostinger’s AI-first approach: What the agentic AI era means for businesses

Instead of waiting for a prompt and returning a response as chatbots do, AI agents can use tools, make decisions, and take actions to work toward a goal.

But how well does that actually work today? Where do AI agents still fall short? And what does an agentic approach look like when you’re actually building and using these systems?

We sat down with Tomas Rasymas, Hostinger’s AI Research Lead, to separate what’s happening now from what’s still further down the road – and look at what the shift toward agentic AI already means for the way we work.

The biggest misconception? That agents can do everything

We asked Tomas what trips companies up first, and he didn’t hesitate: it’s the belief that AI agents can do everything.

They can’t. At least, we’re not there yet.

For an AI agent to actually be useful, Tomas says it needs three things:

• A clear goal. The agent needs to know exactly what it’s supposed to achieve.
• Context. It needs the information required to operate successfully.
• Boundaries. It needs clear limits on what it should and shouldn’t do.

Without these, even a highly capable agent won’t necessarily produce useful results.

That’s why Tomas advises against trying to integrate AI agents into everything. Instead, pick one simple problem and solve it first.

Once you’ve figured out how to make an agent work for that use case, you can iterate, learn from the results, and move on to bigger problems. Starting small gives companies a way to build on what’s working rather than trying to solve everything at once.

A better way to adopt an AI agent is to treat it as you would a new employee. If you’re wondering what that looks like in practice, our guide to maximizing AI agents walks you through how to get the most out of them.

Why should companies take AI seriously now?

AI is already widely used by companies to solve real problems, from recommendation systems and fraud detection to forecasting. But these applications were typically built to solve specific problems within defined parameters.

According to Tomas, we’ve now seen a more significant shift: AI models can work with a range of tools, opening up possibilities beyond predefined applications.

AI isn’t limited to generating an answer anymore. It can automate tasks, make decisions, and take action.

That combination of capable models and tools has made AI look less like a passing trend and more like something companies need to learn how to use.

And in fact, integrating AI into your stack doesn’t equal building a sophisticated system.

Even small businesses can use AI in simple, practical ways to take work off people’s plates, from automating repetitive tasks to handling everyday business processes.

The possibilities vary widely depending on the business and the problem you’re trying to solve. The important thing is to start looking at where AI can do useful work.

What does being AI-first actually look like at Hostinger?

For Tomas, being AI-first means looking for problems AI can genuinely solve better or faster, rather than adding it to every process just for the sake of it.

At Hostinger, that plays out simply: if AI can do it, we look for a way to put that capability to work. If it can’t, there’s no reason to force it into the workflow.

That approach shows up in how we use AI internally and how we build AI into our products for customers.

How we use AI internally

One of the clearest examples of how we integrate AI into our everyday work is the use of DEX, our internal AI assistant. Built by our own developers, DEX is continuously improved with new features to support our work and solve our ever-changing business operations needs.

Alongside that, we have a supportive environment for experimenting with and adopting AI tools to improve our workflows, including for non-technical teams. 

For more advanced use cases, our dedicated AI and automation team helps teams across the company automate processes and find better ways of working.

We’ve also run an internal AI Agents hackathon, where teams across the company explored how agents could solve real problems. Some of those ideas have since moved beyond experimentation and into our everyday workflows.

How we bring AI to customers

With the same approach we take internally, we want to make building and growing an online business more accessible, including for people without extensive technical knowledge or the time to handle everything themselves.

That’s where Tomas’ team focuses their work: exploring what’s possible with AI and finding ways to bring it to our customers.

We started by integrating AI to simplify website creation and VPS hosting management, then expanded it to many other areas, including email marketing, ecommerce, and business tasks management.

As AI capabilities continue to improve, we’re evolving the technology behind our products and bringing more agentic capabilities across Hostinger. Our goal is to create a more seamless experience for customers using Hostinger as a unified platform.

One of the most recent examples is the evolution of Kodee, our first AI chatbot. Initially designed primarily as a customer support assistant, Kodee has taken on a much broader role as its capabilities have grown.

Today, Kodee has evolved into Hostinger Agent, going beyond answering questions to help users with customer support inquiries, business management tasks, and other work they need to get done.

So, as AI gets better at answering questions and taking action, one question follows naturally: 

What happens to the human role as AI takes on more of the work?

As AI agents take on more repetitive work, Tomas believes the human role will increasingly center on judgment in ambiguous situations.

AI can process the information it’s given, but it doesn’t always have the context or experience to know what matters when there isn’t an obvious answer. 

There are things we know, understand, or pick up on without ever explicitly putting them into a system – and that can make a difference when a decision isn’t clear-cut.

At the same time, the way we work with AI is changing. We’re already seeing new ways of working emerge across teams, too, as expectations and needs have evolved. 

Our Customer Success team, for example, now includes AI CX Engineers – people who work behind the scenes to improve how Hostinger Agent responds to our clients. They help fine-tune responses and identify ways to make the experience better for customers using Hostinger Agent.

Even as AI becomes more deeply integrated across the company, we’re still hiring across departments, with people taking on new responsibilities and finding new ways to work alongside increasingly capable AI systems.

Asked what would set the best people apart if everyone had access to capable AI agents, Tomas pointed to two things: asking the right questions and owning the outcome.

In other words, the value of working with AI is knowing what you want to achieve, how to guide the agent toward it, and when to step in and make the call.

The agentic era has just begun

The current state of AI agents is less futuristic than the hype sometimes suggests.

Today’s agents still need clear goals, the right context, and boundaries. They can take on meaningful tasks, but they aren’t universal autonomous assistants that can reliably handle anything you throw at them.

Still, we’re already seeing what happens when AI can take action. 

At Hostinger, that shift is evident in how we build products, troubleshoot issues, and help customers turn ideas into websites.

If you want to start bringing AI into your business operations, stay curious and keep asking where it can solve problems better or faster.

Watch the full interview with Tomas on Hostinger Academy: