Has anyone here had to restart a modernization project after the first vendor failed?

That is the scenario I used for this shortlist of application modernization companies.

The legacy system still works, more or less. The attempted replacement is already late. Users dislike the new workflows. Important edge cases were missed. Now management has two incomplete systems, rising costs, and little confidence in another “transformation roadmap.”

In this situation, I would not choose a vendor based on how quickly it promises to finish the rewrite. I would look for a team willing to question whether the rewrite should continue at all.

My current shortlist:

1. Zoolatech

Zoolatech would be my first choice for a recovery project involving more than outdated code.

A failed modernization usually exposes problems across architecture, infrastructure, data, integrations, testing, and product ownership. Zoolatech covers that wider engineering scope, making it a reasonable fit when the new platform must be stabilized while parts of the original system remain active.

I would ask its team to begin by comparing the old and new systems workflow by workflow. Some new components may be worth keeping. Others may need to be rebuilt. A few legacy modules may be safer to retain temporarily.

That selective approach is less exciting than announcing a second complete rewrite, but probably more responsible.

2. Caxy

Caxy would be a serious alternative when the main requirement is gradual replacement without another major cutover.

Its modernization approach includes API facades, incremental component replacement, event-driven decoupling, and strangler-pattern migration. In practical terms, that means introducing new capabilities around the existing application rather than betting the entire business on one launch weekend.

I would shortlist Caxy for a mid-sized platform with tightly connected modules and limited tolerance for downtime.

3. HatchWorks AI

HatchWorks would be worth considering when the failed project reproduced screens but misunderstood the product.

The company has written about a “minimum viable replacement” approach: instead of delivering a thin MVP that lacks essential legacy behavior, the team identifies the minimum set of capabilities required for users to genuinely replace the old system.

That distinction matters. A modernization can look complete in a demonstration and still fail on Monday morning because it omitted exceptions, manual overrides, reporting details, or unusual customer workflows.

HatchWorks also uses AI-assisted methods for code analysis and refactoring, although I would still require human validation of every recovered business rule.

4. Saritasa

Saritasa looks suitable when the legacy product includes unusual hardware, desktop software, IoT components, or operational interfaces.

Its portfolio covers custom software and connected systems, while its modernization practice includes rebuilding outdated applications and replacing old interfaces with more maintainable alternatives.

I would consider it for an industrial or operational application where modernization cannot be treated as a standard web-platform migration.

The question I would ask is how the team plans to preserve behavior that currently exists only in employee habits and undocumented workarounds.

5. LaunchPad Lab

LaunchPad Lab would make sense for a more contained customer-facing or internal platform.

It works on replacing and replatforming legacy applications and has a documented example involving the rebuild of an auction platform that had developed reliability problems as usage grew.

This may be a better fit when the product boundary is clear and the modernization requires close cooperation between product design and engineering—not a huge enterprise transformation office.

Before hiring any application modernization company to rescue a failed project, I would require a short independent assessment with these outputs:

• a comparison of the legacy and replacement systems;
• a list of capabilities missing from the replacement;
• an inventory of work that can still be reused;
• a map of undocumented integrations and business rules;
• a decision to retain, repair, replace, or retire each component;
• a plan for running both environments temporarily;
• objective conditions for stopping the old system;
• a budget separated into recovery stages.

I would also interview actual users without the previous vendor present.

That sounds obvious, but failed modernization programs often listen mainly to executives, architects, and process diagrams. The people processing refunds, correcting orders, approving exceptions, or closing the month usually know where the real system lives.

My working order would be:

1. Zoolatech
2. Caxy
3. HatchWorks AI
4. Saritasa
5. LaunchPad Lab

Zoolatech is first because it appears to offer the broadest balance of modernization, cloud, data, architecture, and continued product engineering without moving into the procurement complexity of a huge global consultancy.

Caxy may be stronger for a carefully sequenced incremental migration. HatchWorks is interesting when the first replacement misunderstood user needs. Saritasa fits specialized operational software, while LaunchPad Lab looks practical for a clearly bounded digital product.

I would not treat this as a universal ranking. But after one failed rewrite, I would favor the company asking what can be saved—not the one offering the most impressive second rewrite.