ReCore
Also: Legacy-to-AI Modernization program
What is ReCore?
ReCore is our eight-to-sixteen-week Legacy-to-AI Modernization program, priced from $31,500 to $105,000 and above, that modernises a legacy system incrementally and adds AI capability without stopping the business.
What ReCore means
ReCore is the program we run when the system that matters most is old: a PHP 5 ERP, a monolith with no API, a university administration platform nobody dares touch. Over eight to sixteen weeks we put a safety net of characterisation tests around the existing behaviour, add an API layer, and replace or extend modules using the strangler pattern so the old system keeps running throughout.
AI enters where it removes real work: document intake, reconciliation, natural-language reporting, agents that handle exceptions. We do not bolt a chatbot onto a broken data model. Price ranges from $31,500 to $105,000 and above, set after a discovery phase that maps the system and orders the modules by risk and value. Each phase has its own zero-downtime cutover plan and a hypercare period.
ReCore is not a rewrite. A rewrite freezes the business for a year and usually fails to reproduce the undocumented rules that keep the company running. It is also not a lift-and-shift: moving an old system to new servers changes the bill, not the problem.
Who it really matters to
- CTO / Head of Engineering: you get a documented, tested and API-accessible system without the risk of a big-bang replacement.
- Operations head: the business keeps running through every phase, with cutovers planned around your calendar.
- CFO: the program is phased and fixed-price per phase, so spend follows delivered value rather than a single large bet.
- Founder / CEO: the legacy system stops being the reason you cannot ship AI features or integrate with partners.
- Compliance officer: modernisation is the moment to add audit logs, access control and data-handling rules that the old system lacked.
Why it exists
ReCore exists because most legacy modernisation projects fail in one of two ways: a full rewrite that never ships, or an AI layer stuck on top of a system whose data cannot support it. Incremental modernisation with tests and a strangler pattern is slower to describe but far more likely to finish. The trade-off is patience: the old and new systems coexist for a while, and some technical debt is carried deliberately until its module is reached. We accept that because a business that keeps operating through modernisation is the only kind that gets to the end of it.
Where it is applied
- A university modernising its student records ERP module by module around the academic calendar.
- A manufacturer replacing a custom PHP ERP's inventory and invoicing modules while adding GST e-invoicing integration.
- An NBFC exposing its core lending system through an API layer so AI document intake can feed it without re-keying.
- A logistics company modernising a fleet and settlements platform and adding exception-handling agents.
- A hospital replacing a legacy admissions system with one that supports consent tracking and ABDM integration.
Is ReCore a skill?
Eazyware programAn Eazyware program: eight to sixteen weeks, fixed price per phase from $31,500 to $105,000 and above, delivered under the Legacy-to-AI Modernization service. It combines modernisation engineering with AI capability, and hands over to a Care Plan for ongoing support.
Eazyware service that covers it: Legacy-to-AI Modernization Program. Starting prices are on the pricing page.
Frequently asked questions
How do you set the price for a ReCore engagement?
A discovery phase maps the system, its integrations and its undocumented rules, then orders modules by risk and value. Each phase is then quoted at a fixed price with a fixed cutover date, so the total is built up from parts you can see.
Will the business have to stop during modernisation?
No. The strangler pattern keeps the old system live while new modules take over one at a time, and each cutover is planned for zero downtime with a rollback path and a hypercare period afterwards.