Skip to main content
Own your model starts with definable blocks: a block takes communication in and writes structure out - extraction, qualification & scoring, summarization, sentiment & intent, evaluation, or a custom block you define. Today blocks run on call outcomes; where the system is headed is placement anywhere in the pipeline - attached to a single step, a conversation, or the whole stream (in build). Every insight lands back in the Context Engine, where the next block - and the next agent - can read it. That loop is where this goes. Because outcomes accumulate in one durable system of record, the memory can become a model - one only you can train, because no one else has your conversations. The first two steps run in production today; the rest is planned.
1

Blocks run on every interaction

Extraction, scoring and summary blocks turn each call into structure. Production
2

Outcomes land in state

Every completed call writes a structured outcome you can read back. Production
3

A dataset builder turns history into training data

Planned
4

RL simulation plays the workflow against caller personas

Planned
5

A per-org SLM (~1B) is trained on your outcomes

Planned

​
The primitives

​
On the spine

Blocks read interactions, outcomes, and context; they never touch the audio path. What a block produces is written back to the spine - onto the outcome it analyzed or into the Context Engine - so downstream blocks and agents read it as ordinary state, not as a side channel. Each insight written emits an event, like any other change on the spine.

​
What stays hidden

How a block is prompted and scored. You define blocks and ratify evaluations; how a block runs is not a surface you configure. When the training pipeline ships, its internals stay hidden the same way.

​
Status

  • Definable blocks: extraction, scoring, and summarization run on call outcomes in production today via Post-call data; placement on arbitrary steps and streams is in build.
  • Evaluations: conversation-level evals callable today through SuperDialog evals.
  • Per-org SLM training (dataset builder, RL simulation, trained models): planned - not available yet.

​
Go deeper

Post-call data

The API where extraction, scoring, and summarization results land today.

Outcomes

The structured records blocks read from and write back to.

Context Engine

Where insights persist as retrievable memory.

SuperDialog evals

Score conversations against ratified expectations.