WPAICG (AI Engine by AIPower) bundles a chatbot, a content writer, image generation and an embeddings store into one plugin. Each subsystem writes its own log: chatbot sessions and messages in wpaicg_chatlogs, content writer runs in wpaicg_writer_logs, embedding queries in wpaicg_embeddings. The plugin's admin gives each subsystem its own tab.

That works for inspecting one subsystem at a time. It does not work when the operational question is cross-subsystem: which user is responsible for most of the activity, which model is doing the work this week, what the total spend across chat, writer and embeddings looks like.

SleekView reads the three log tables into a single unioned dataset where subsystem becomes a column. Sort by tokens_used to see the heaviest rows regardless of where they came from. Filter to a single subsystem when you need that view, or filter across all three for a cost-and-usage audit. Capability gates keep message bodies and writer output scoped to the right role.

WPAICG bundles four AI surfaces into one plugin, and bundles four separate admin tabs to go with them. That is fine when you want to inspect one subsystem at a time and inconvenient when the operational question crosses surfaces. A unioned table makes subsystem a column, which means sort-by-tokens, filter-by-user and group-by-model all work across chat, writer and embeddings at once.

Capability gating keeps message bodies and writer output scoped to the right roles, so cost reviews can stay open without exposing conversation text. Same logs, same plugin, completely different posture toward cost, growth and per-user accountability.