Introduction

As 2025 came to a close, we did what we always do at year-end: looking back, taking stock, and asking ourselves what actually mattered.

The answer? A lot.

This year didn't just bring projects. It brought momentum. Growth. A level of complexity we hadn't operated at before. And somewhere between the rebrand in the beginning of 2025, the office move in summer, and the wave of work that carried us through autumn, we became a different version of ourselves.

We launched new products. We built tools we now use daily. We helped companies access funding. We automated a lot of tasks. We wrote over 50 blogs about AI, reaching hundreds of thousands of readers, establishing Solveo as a voice in the space we're helping define.

And we did it as a team that's grown, celebrated anniversaries together, drank way too much coffee, laughed through hard work, and learned a lot! 

This is the story of how we closed 2025. Not just with completed projects, but with work that's carrying forward into 2026. Work that's teaching us what we're capable of. Work that's redefining what Solveo can be.

The projects that pushed us to think differently

Working with One of the Fastest-Growing AI Coding Tools

This one started with a simple question: what if developers actually liked their coding tool?

In 2025, we began collaborating with one of the fastest-growing AI coding assistants. A tool that helps developers build, ship, deploy, and iterate faster. Plan code. Write it. Debug it. Move faster between ideas and working features.

It works with multiple models, lets developers bring their own API keys, and has different modes depending on what they're doing, planning, building, fixing things. Simple, practical, and very usable.

What we worked on together:

We spent the second half of 2025 in Reddit threads, GitHub discussions, dev communities. Not promoting. Just being there. Listening to what frustrated people, what made them switch tools, what they actually needed.

We:

• Built a full Reddit engagement system
• Took part in 1,000+ Reddit discussions across 60+ subreddits
• Distributed content that reached ~500,000 readers, including ~336K views from Reddit alone
• Achieved record engagement levels and became part of the daily conversation in key dev communities
• Supported design improvements

This wasn't a one-off campaign. It was ongoing work, side by side.

Then we did something different: we started using it ourselves. Built internal tools, small automations, experiments that made us faster. We weren't selling the product anymore, we were living with it. Testing it like our own work depended on it, because it did.

And then, not just for fun. We used it to build one of our biggest projects this year: Funding.mk.

In the middle of all this, they raised $8 million in investments.

Watching a product you work closely with grow like that feels really good. We're genuinely happy for the whole team.

What we loved most: the moment we stopped talking about it and started talking with it. The product grew because we genuinely knew it worked, not because we claimed it did.

This collaboration helped their team grow. And it helped us learn how to build with AI, not just talk about it.

When you truly like the product you're working on, everything flows better. Can't wait to bring more ideas to the table in 2026 and see where they go next.

Funding.mk: Turning Access to Finance into Reality

Many SMEs struggle to identify and secure suitable funding for digital transformation and innovation. They lack the internal capacity to navigate complex EU and national funding programmes.

Through INNOFEIT EDIH, Solveo provided hands-on advisory support, helping SMEs map relevant opportunities, structure project ideas, and prepare high-quality funding applications tailored to their growth and digitalisation needs.

The service combines innovation consulting with access-to-finance expertise, turning abstract digitalisation goals into concrete, investment-ready project pipelines for local companies.

And we built the platform with the AI coding tool we mentioned above.

The result? Developing and submitting over €10,000,000 in funding applications across EU and national programmes, increasing companies' chances to attract external finance for digital innovation projects.

Funding.mk became more than a project. It became proof that when you combine the right tools with the right expertise, you can move fast and build something that actually works.

Expanding Audience Discovery for a Global Training Provider

Earlier this year, we started working with a global professional training provider facing an interesting constraint: they had sophisticated marketing systems running at full capacity, but they'd hit a ceiling. LinkedIn targeting, paid search, CRM enrichment, all optimised, all working, but maxing out on who they could reach.

The challenge wasn't about doing more of what worked. It was about finding a way to reach professionals who simply didn't exist in their current systems.

So we built something that listens instead of targets.

A real-time audience intelligence system that pulls signals from places like X (Twitter), skills mentioned in bios, industries discussed in threads, roles revealed through conversations. Not scraping profiles, but understanding patterns. The system identifies, enriches, and prioritizes professionals who match specific criteria: skills, roles, industries, regions. Then it delivers them as structured, actionable leads in seconds.

What made this work: treating social behavior as market intelligence. People naturally signal intent and expertise when they're not being marketed to. We just built the infrastructure to catch it.

The result? A continuous stream of high-relevance professionals who literally don't exist in any other channel they use. Fresh leads, zero overlap with existing sources, same quality bar. No saturation risk. No competing with their own paid campaigns for the same audience.

They can now identify thousands of relevant professionals in real time—expanding not just their reach, but their actual capacity to grow.

AI-Driven Skills Assessment & Learning Recommendations

With the same client, we tackled a different kind of navigation problem.

They had 1,200+ courses across dozens of industries. Everything a learner or organisation might need, but that was part of the problem. Where do you even start? Which skills actually matter? What should companies prioritise when upskilling teams?

Traditional course recommendation systems weren't built for this. They suggest what's popular or adjacent to what you've done. They don't tell you what you're missing or what matters most for where you're trying to go.

We built an AI skills engine that reads between the lines of your experience.

Upload a CV or LinkedIn profile. The system analyzes 1,000+ capabilities, compares them against real workforce demand, identifies the gaps that actually matter, and generates a full upskilling roadmap, prioritized, personalized, instantly actionable.

Not here are some courses you might like. But here's what you need, why it matters, and where to start.

What made this work: treating skills as a system, not a list. Understanding how capabilities cluster, which gaps compound, and what learning sequences actually make sense in practice.

The catalog became a strategic tool. Learners get precise guidance instead of endless browsing. Companies can map skill gaps across entire teams and prioritize upskilling investments based on actual data, not guesswork.

Two different problems, audience discovery and learning navigation, but the same underlying principle: good systems don't just recommend. They orient. They expand what's possible, not just what's convenient.

We're looking forward to what we'll build together in 2026. We'll probably talk more about it then.

Automating Thousands of Calls with AI Voice Agents

As we wrap up 2025, we're reflecting on the projects that pushed us to think differently. This one was all about slowing down to get it right.

A service organisation (under NDA) managing thousands of follow-up and reminder calls every month wanted to reduce the operational burden of routine conversations.

We're developing an AI-driven voice agent capable of conducting structured outbound conversations end-to-end, including information validation, rule-based decision-making, and follow-up flows.

We didn't jump straight to building. Instead, we spent time mapping the real process, identifying risks, and designing an architecture that would actually work in production. We built conversational flows in Macedonian, determined when the AI should hand off to a human, and ensured compliance with legal and privacy rules.

What we loved most: the discipline it took to design something that works with people, not around them. AI where it helps. Humans where it matters.

Next up is a controlled pilot to validate everything before scaling.

It converts a high-volume, repetitive operational process into a fully automated system, improving efficiency, freeing teams for complex tasks, and maintaining consistent communication quality at scale.

What 2025 taught us

Looking back, 2025 wasn't just about doing more. It was about doing different.

We learned that scale doesn't have to mean losing yourself. Our team grew. We celebrated anniversaries together. We had celebrations outside the office. We had fun. We grew together while working harder than we ever have.

We learned that the best work comes from actually liking what you're building. The projects we loved most were the ones where we genuinely believed in what we were making.

We learned that momentum builds quietly. It's not always the big announcements or the flashy launches. Sometimes it's just showing up, doing the work, and realizing months later that you've changed without noticing. 

What's next for 2026?

We're not stopping.

The projects we started in late 2025 are carrying forward. The AI coding tool partnership continues. The voice agent moves to pilot. The training provider's work expands. Funding.mk keeps growing. And in the meantime, we started a few (pretty interesting projects) that we will tell you about another time… 

But beyond that, we're curious. About what AI can do next. About what our team can build. About what problems we haven't solved yet.

2025 proved we could handle scale, complexity, and the kind of work that stretches you in unexpected ways.

2026? Let's see what happens when we stop being surprised by our own capacity.

So, yes… this was our 2025 wrapped. Thanks for following along.

If you missed the earlier parts of this series:

• How 2025 Started: Rebranding, Redefining, and Building New Foundations
• Part 2: The season of momentum (Building, Testing, and Organising)

Partner with Solveo!

Implement change. Automate smarter with AI. Be on top of global trends.

Get in touch. 

‍