Mechanism chain in four beats:

1. Cells from a positive blood culture are exposed to antibiotics in the assay.
2. A microscope captures time-lapse video of individual cells.
3. Computer vision models score response patterns linked to susceptibility vs resistance.
4. The system reports a categorical interpretation for listed drugs.

LifeForce owns the infection / right-drug timing outcome. AI explains the capability that makes same-shift reads feasible. Invent must not become “build a neural net that invents antibiotics.” Hand players faster right-drug answers after positivity, not an ML bake-off.

Vision does not remove blood culture positivity, Gram-stain constraints, separate organism ID entry, prescription-use context, or stewardship judgment after the report.