Sense-Reason-Act-Learn

Definition. The atomic four-step cycle an autonomous GTM agent runs: sense signals (zero/first/third-party data, intent), reason over them against a knowledge base and policies, act across channels and systems, and learn from outcomes and human corrections. The unit that populates the execution tier — distinct from rule-based automation, which only triggers preset actions.

@misc{shalvi_gtm_sense_reason_act_learn,
  author = {Singh, Shalvi},
  title  = {Sense-Reason-Act-Learn — GTM World Model Glossary},
  year   = {2026},
  url     = {https://shalvisingh.com/gtm/glossary/sense-reason-act-learn}
}