Your whole stack becomes Duplex’s toolset.
RAG, caches, knowledge graphs, search, databases, APIs and models: Duplex calls each one when a step needs it, and learns which works best.
One toolset. Nothing replaced.
It learns which tool to call.
Confirmed results decide which tool each kind of request gets, and keep deciding as your sources change.
For a new kind of request, Duplex tries the tools that could help.
The difference learning makes.
The same tools on their own, against Duplex, on the same work.
After a rule changed, a cache fell to 42.6%. Duplex held 92.5%.*
Operational intelligence.
Less repeated work, and deeper reasoning where it pays.
Your retrieval, caches, knowledge graphs, APIs and models become tools Duplex can call.