> ## Documentation Index > Fetch the complete documentation index at: https://docs.unpod.ai/llms.txt > Use this file to discover all available pages before exploring further. # Outcomes > The structured result of every interaction - not a log entry, the actual condition the next step reads. An outcome is the structured result of an interaction. Not a log entry - the actual condition the next step reads. When a call ends, a message gets a reply, or an email bounces, the interaction resolves into typed fields that the rest of the system acts on. The rule the whole Execution layer is built on: **the outcome of step N is the input of step N+1.** The system defines a small set of default fields centrally and keeps it deliberately few - roughly five: connection (did we reach them), interest, callback (when to try again), channel preference, and disposition. On top of those, companies define their own custom outcomes - exactly as they already can for calls today, where post-call structured data carries whatever schema the use case needs. One class of outcome is special: an **exit outcome** removes a contact from a journey entirely, so "do not call again" is a state the engine enforces, not a note someone hopes gets read. ## The primitives | Primitive | What it is | Where it lives today | | ---------------------- | ----------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------- | | Outcomes | The structured, versioned result of one interaction | Production for calls - delivered as [post-call data](/api/post-call-data) | | Outcome fields | The centrally defined defaults: connection, interest, callback, channel preference, disposition | Extracted on every call today; see [post-call data](/api/post-call-data) | | Custom outcome schemas | Company-defined fields layered on the defaults | Defined per agent for calls today via [post-call data](/api/post-call-data); journey-wide schemas in build | ## On the spine ``` INTERACTION -> OUTCOME ---> EVENTS | | +-- CONTEXT ``` Outcomes read the finished interaction - transcript, turns, tool results - and write one structured record against it, which rolls up into the conversation and feeds the entity's context. Flows and journeys read that record to pick the next step; an exit outcome ends the journey for that contact. Every recorded outcome emits an event: today that is the post-call webhook, delivered per [Hooks & Events](/speech-stack/hooks-events). ## What stays hidden Extraction internals. How an outcome is pulled from a transcript - which models run, how fields are validated against the schema, how versions are reconciled - is not a public surface. You define the schema and read the result; the extraction pipeline between them is Unpod's to run and improve without breaking your fields. ## Status * **Production** for calls: post-call structured data is how calling reports today - every completed call resolves to an outcome you can receive by webhook or fetch by API. * **In build**: journey-level outcome branching - the journey engine reading outcomes to choose the next step across channels. ## Go deeper Receive call outcomes in your system today - transcripts, summaries, and extracted fields by webhook. React to call lifecycle events, including the post-call moment where outcomes land. The engine that reads outcomes to decide step N+1. The definable blocks - extraction, scoring, summarization - that turn an interaction into an outcome.