AGENTIC WORKFLOWS
The conversation flow your ministry needs, no code required.
Evaluators read where a seeker actually is in their journey. Routing decides what happens next — a resource, a follow-up, or a handoff to a real person.
01
Evaluators notice what a reply alone can’t
An Evaluator judges a piece of text and reports a verdict — a category (such as Positive, Negative, or Neutral) or a 0–1 score against criteria you write.
That lets a ministry flag distress and offer a human handoff instead of another paragraph, catch a hostile or bait prompt before an answer goes out, tag every conversation by topic in the background, or score answer quality before switching models.
02
From verdict to action
The same verdict is read differently depending on where it’s used. A Call to Action shows content or offers a handoff when the verdict matches. A Custom Guardrail blocks, warns, or redirects before or after a response. An Automation applies tags and metadata in the background. A Benchmark scores an agent’s answers across a whole question bank.
03
Voice agents
Let seekers speak instead of type. Choose a voice persona and model, and the agent answers out loud in the language the seeker is using — grounded in the same sources and guardrails as its text replies.
01
Create Conversational Pipelines
02
Save User State w/o Compromising Their Identity
03