Ask any question about your FX data. Get an instant answer.
Knox is Kantox’s AI-powered FX assistant — grounded in over a decade of FX risk management expertise — allowing treasury teams to get precise, plain-language answers on orders, positions, and platform questions the moment they need them.

The answer already exists.
It's just now instantly reachable.

Instant answers grounded on expertise

FX native

Grounded on real data

Instant precise answers

No wasted time
Privacy and security by design
Knox delivers instant, precise answers straight from your own FX data

Your data is never sent to, stored or trained by a third party. It stays inside Kantox

Account numbers, bank codes and other sensitive fields are stripped out before Knox starts to work

Every answer is traceable. Knox is built to meet EU AI Act requirements and never gives legal or financial advice.
Book a demo and see what your own FX data can tell you
FAQs
While a Multi Context Protocol server (MCP) would give an AI assistant access to tools to query your data, it would lack a key ingredient: context. Knox, on the other hand, is purpose-built for FX risk management from the get-go. Knox encodes a company’s hedging policy document. This reduces ambiguity and overrides manual intervention. Your own hedge and trading history lives inside Knox, and reporting comes naturally as a byproduct of using the platform.
No. Knox is read-only by design — it explains and shows, it never acts on your behalf.
No. Your data is never used to train models and never leaves Kantox's infrastructure. Only aggregated or redacted data reaches the AI model.
No. Every query is scoped to your own company, enforced at the platform level.
Anything about your own FX orders, settlements, and positions, plus how-to and conceptual questions from a curated knowledge base. Program-level KPI reporting (WAHR, hedge ratio dashboards, etc.) is on the roadmap.
No — Knox is grounded in Kantox's execution-level data and more than a decade of FX risk management expertise, not general-purpose model knowledge.