We modernize legacy systems your business runs on.

Changing the wheels on a moving car. Modernizing critical systems for mid-market companies without breaking the business.

Book a 30-min call
01 Problem

A legacy system that can no longer keep up with the business

Most companies with a legacy system know the same story. Maintenance gets more expensive. A new feature takes months instead of weeks. Connecting the system to modern data, analytics, or AI isn't really feasible. The people who originally wrote it are often no longer with the company.

Rewriting the whole thing is expensive and risky. Leaving it alone means slowly tightening the noose: costs grow, the people who understand the system dwindle, and the business loses room to move.

Between a full rewrite and stagnation, there is a third path. Modernizing the system with a plan and in pieces so operations don't stop and the investment doesn't go to waste a year later. That's the path we walk with you.

02 Who it's for

For CTOs, CIOs, and CEOs of mid-market companies

Where:

  • in-house software powers the business (logistics, energy, manufacturing, finance, e-commerce)
  • the system can't keep up with the speed of the business, either because it's more than five years old, or because the company is growing faster than development can match
  • after an acquisition, systems need to be integrated or rewritten onto a unified stack
  • there is pressure (from inside and above) for AI and data, but the architecture can't deliver
  • modernization can't wait, but no one wants to risk a rip-and-replace

If any of this sounds like your situation, you're in the right place.

03 How we work

Modernization stands on architecture

Before the first line of code, we design the target state: data model, module boundaries, integration points, data layer. This is where it gets decided whether in three years you will have a modern system, or just a newly wrapped old problem. Most failed modernizations fail here, not in the code.

Project phases

  1. 1
    Analysiswhat the system actually does, not what it was supposed to do according to the documentation
  2. 2
    Target architecturedata model, modules, integrations, a data layer ready for AI and other systems
  3. 3
    Developmentwe choose approaches that remove technical debt for the long term, not just the most visible parts. We document the solutions we build carefully, because that is what keeps the system clean and easy to extend. Newly discovered legacy issues are regularly reviewed with analysis and users, and we don't migrate bugs
  4. 4
    Testingfor rewrites we verify parity with the old system, often against real production load (shadow traffic, parallel runs, output comparison)
  5. 5
    Deploymentwhere it makes sense, we deploy gradually by sprint and phase. For smaller or isolated systems we sometimes go bigbang, when that's a safer path than long parallel operation. What is right comes from the architecture and the risk, not from a template
  6. 6
    Iterationafter go-live we don't drive off. Tuning, additional modules, gradual extension

AI in our work and in your system

We use AI heavily in modernization ourselves and have our own approaches and methodology for it. The goal is twofold. Your new system has the architecture, data model, and APIs on which it makes sense to build an AI layer. And it's set up so that effective AI-assisted development is possible going forward, which means another major rewrite won't be needed in three years.

To be honest about it: AI opens new paths, but modernization still requires plenty of work and expertise. Where deterministic code makes sense, we use deterministic code. Where an LLM makes sense, we use an LLM. The value lies in knowing which is which. We can build most basic AI use cases ourselves (AI agents on top of LLMs); more advanced cases we tackle with partner AI firms (e.g., inventory optimization, e-commerce recommendation, energy forecasting).

We start modernization projects with a structured Assessment that diagnoses the system and proposes concrete paths forward. We'll discuss the details on a call.

12+ years
building critical software for mid-market and enterprise
5,000+
MDs invested in our largest rewrite project
45+
successfully delivered projects, short- and long-term (not only rewrites)
1 year
from kickoff to production on our flagship rewrite
04 References

Modernizations running in production

E-commerce / logistics

Mall Group — Delivery platform rewrite

For Mall Group (part of Allegro Group since 2022) we rewrote the entire logistics and delivery solution for the whole group, including the system for the central warehouse in Jirny and the SAP integration. It was a critical system covering shipping calculation in the basket, delivery management, and warehouse coordination together with their ERP/WMS in SAP. Without it, nothing would have sold or shipped on time. Our analysis ran all the way from warehouse processes down to jobs running inside SAP. We brought the project to production within one year and launched it right before the Christmas season. The deployment was an exceptional success: Mall became the technology leader in delivery planning, a level other large e-shops did not match for a long time. The system today serves 100+ carriers and thousands of marketplace partners.

This was preceded by two smaller projects for Mall Group on an old Groovy stack in a technically unsustainable state: first a rewrite of the logistics solution for acquired e-shops, and then a rewrite of the financial module. The motivation was integrating acquired e-shops into a single system. As part of the financial module, we used vertical slicing out of the monolith to extract payment processing and built a modern platform for payment matching and financial distribution across the e-shops.

Asset management / .NET

Large-scale CMMS system (long-term, 5,000+ MDs)

Modernization of a 20+ year-old asset management system: 700 forms, 10+ modules in a monolithic architecture. The system ran on an unsupported .NET version, with third-party components from defunct vendors and spaghetti code across all layers down to stored procedures. We initially considered vertical slicing into microservices, but after a thorough analysis we chose horizontal layer separation and a gradual in-place rewrite. The first phase is in production: legacy webforms removed, the system runs on supported technologies, with no business downtime. We continue further.

Under NDA, we do not disclose the client's name or further details.

Auction platform

OK Dražby — Auction portal rewrite

A complete rewrite of an auction portal that couldn't handle more than 10 concurrent auctions and was blocking new feature development. In this case we chose bigbang over a gradual transition: the scope was small, the data structure simple, the technology stack unsuitable for gradual migration, and development on the old system could be frozen for the duration of the project. Bigbang makes sense where the cost of gradual migration significantly outweighs the risk of a large deployment. Here we invested heavily in migration risk mitigation, performance testing, and thorough analysis.

The migration ran over a weekend, for tens of thousands of users. The deployment went through with virtually no downtime; aside from one bug affecting a single auction (out of a thousand), there was no other significant problem. We fixed the bug immediately. The system has been in production for over a year, and we keep extending it.

Read the full case study →
Industrial IoT

Mecc Alte — SmartCloud

For an Italian manufacturer of industrial generators we built SmartCloud, a new IoT platform for real-time device management and monitoring. The original solution was slow, didn't scale, and behaved unstably. Instead of an in-flight rewrite, we chose to build the new system alongside the existing one and migrate clients onto the new platform gradually. This approach makes sense where the technology and architecture differ so much that a piecewise rewrite is not realistically feasible. It also helps that this is a product where clients don't have to be migrated on a single day (which is not true for many companies' core systems).

HVAC / predictive maintenance

Sensible — SmartCoil

A US client came to us with SmartCoil, a platform for predictive HVAC maintenance built for them by a local US firm. The system was in poor technical shape, lacked proper architecture, and parts of the functionality were unusable in production or worked with serious bugs. It wasn't a classic in-flight rewrite, but the starting point was identical to a legacy modernization. We fixed fundamental flaws at the foundation while adding new features in parallel. We are still working with the client today; they are happy, and thanks to the modernization we are building new features that support their expansion.

“
Every project with Coding Bear started with a thorough analysis of our situation and needs, followed by a clear architecture proposal, and only then development and integration into our existing systems. Our experience was smooth, efficient, and professional. I wholeheartedly recommend them.

Petr Mahdal, then CIO at Mall Group

Worth a quick call?

If this page describes the situation you are in, it makes sense to spend 30 minutes on a call.

It's not a sales call. A short diagnosis: where you are, what the biggest risk is, whether it makes sense for us to keep talking. If it does, we'll send you the details of how we actually kick off modernization. If not, we'll say so on the call.

Book a call