Inside NZC AI and ESG AI
The layer already sits behind the standalone products. Consultant and client tools continue alongside it — the layer is an addition, not a replacement.
MCP layer
The intelligence layer that can sit inside other software
The MCP layer is the intelligence layer across our software products: in-product assistance, document ingestion, data validation, deterministic calculation and continuous compliance updating. It is separately licensable to third-party platforms.
Capability
What the layer provides
The same layer that supports our own products can support a partner's product, with the host platform keeping its own interface and workflow.
In-product assistant
An assistant inside the host product that explains functions, answers questions about the screen a user is on and guides them through a task.
Document reading and ingestion
Uploaded evidence is read and used to populate fields, so information is captured from the documents a user already holds rather than typed twice.
Extraction, validation and correction
Extracted data is checked against expected ranges, units and formats, with issues flagged and corrections proposed rather than applied silently.
Calculation and compliance updates
The deterministic calculation layer and the compliance methodology behind the host product, maintained against UK and EU regulatory change.
Development and fault-finding support
Development, testing and fault-finding support for the platforms the layer serves, so issues are traced against real data and methodology.
Licensing
How the layer is used
Licensed to a third-party platform
Software providers can license the layer so their own product gains assistance, ingestion, validation and compliance calculation without building it.
Alongside NZC Portal
Within the wider NZC Portal workflow the layer supplies the specialist intelligence, with portfolio and building context available to it.
Licensing scope, integration route and commercial terms are agreed per platform. Speak to us for current terms.
Assurance
How output is produced
Numerical output in the layer is produced by deterministic calculation engines and is never generated by a language model. Compliance methodology is maintained continuously, and reports issued through a host platform remain subject to qualified professional sign-off.