Short answer
Yes. RAG and retrieval is a core capability of EngineAI, the build team of FirstCoreAI, FirstNet's AI business unit. EngineAI makes documents, contracts and internal knowledge searchable and answerable in plain language, and takes the system from proof of concept to production.
In detail
Retrieval-augmented generation (RAG) finds the passages in your own content that are relevant to a question and passes them to a language model, so answers draw on your material rather than only on what the model learned in training.
How EngineAI approaches it:
- InfraAI, the AI Factory service, suits RAG and retrieval back-ends, with multilingual embeddings and a retrieval reranker available through OpenAI-compatible endpoints
- Most EngineAI projects run on InfraAI compute in FirstNet's South African data centre, so documents, models and pipelines stay in the country, or the system can be deployed in your own environment
- Open-weight models are the default, with frontier APIs only where justified
- Integration connects the solution to Microsoft 365, your CRM and other line-of-business systems
FirstCoreAI benchmarks every model on the AI Factory against frontier APIs for accuracy, latency, concurrency and cost per unit of work, so you see the evidence before committing. FirstCoreAI is already in production in financial services and insurance, where claims and document extraction are typical workloads.
A FirstCoreAI scoping call is the place to begin.
Source: FirstNet Custom AI Engineering service page →
Didn’t answer your question?