You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Proposal: a clarification-loop example for decision models with incomplete input
#4608
I'm Farouk, CTO of Zelinqa. We develop a conversational questioning engine that tracks collected information and selects the next question from a configured domain.
I'd like to propose a small integration example demonstrating how decision models can work with an upstream information-gathering layer.
Example workflow
A user describes their requirements.
A Decision model evaluates a routing decision.
If required information is missing, the application requests a relevant clarification question.
The user's answer updates the state, and the decision is evaluated again.
For example, a request contains a clear need and timeline, but no budget. Rather than immediately selecting a sales route, the application asks about the budget first.
Missing information would be tracked explicitly: a probability near 0.5 alone would not establish that information is missing.
The proposed example would keep the questioning layer separate from model inference, with Zelinqa as one possible implementation—not a required dependency of Semantic Router.
It would use synthetic conversations and compare decision-only versus clarification-assisted workflows, measuring decision accuracy, additional turns and end-to-end latency.
This is currently an integration proposal, not a tested Decision-model integration or a benchmark claim.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Hi everyone,
I'm Farouk, CTO of Zelinqa. We develop a conversational questioning engine that tracks collected information and selects the next question from a configured domain.
I'd like to propose a small integration example demonstrating how decision models can work with an upstream information-gathering layer.
Example workflow
For example, a request contains a clear need and timeline, but no budget. Rather than immediately selecting a sales route, the application asks about the budget first.
Missing information would be tracked explicitly: a probability near 0.5 alone would not establish that information is missing.
The proposed example would keep the questioning layer separate from model inference, with Zelinqa as one possible implementation—not a required dependency of Semantic Router.
It would use synthetic conversations and compare decision-only versus clarification-assisted workflows, measuring decision accuracy, additional turns and end-to-end latency.
This is currently an integration proposal, not a tested Decision-model integration or a benchmark claim.
I describe the idea and an illustrative Jev workflow here:
https://zelinqa.ai/en/engineering/decision-models-next-question
Would this fit a community example or recipe? Is there an existing effort we should coordinate with?
All reactions