Comparison tables gain a narrative axis judging whether each entity's
community conversation is about what the entity pitches or about
something else (pricing, rivals, an incident). Backed by a new
mandatory Step 0.55 research item that fetches each entity's CURRENT
first-party positioning (RESOLVED_POSITIONING) instead of relying on
memory, and a narrative-check synthesis beat for single-entity company
runs. The mismatch is the signal: companies usually don't control
their own conversation.