TEXT BEHAVIOURAL SIGNALS / NYETI API
Nyeti turns customer messages into five behavioural signals and response guidance for support teams and AI workflows.
Nyeti evaluates customer support conversations, so you can see where the resolution didn't actually land.
Closed beta. Applications are reviewed before access is offered.
Incoming message
I need this fixed today. Your bot is useless and I want to cancel.
- Emotional toneNegative
- Mental effortNeutral
- UrgencyHigh
- EngagementDisengaged
- Communication styleNeutral
Response guidance
- Posture
- Empathize and solve
- Tone
- Warm and direct
- Action
- Acknowledge, then fix
- Input
- Customer text
- Signals
- Five dimensions
- Output
- Response guidance
- Control
- Your workflow decides
Conversation context
Follow behaviour across the conversation
Each turn uses the same analyse call. Add conversation_id and speaker when conversation context is enabled.
Nyeti can then flag repetition, false resolution, and escalation. Single-message analysis still works on its own.
See conversation contextconversation_id: ticket_123- Customer / 01
My order still has not arrived.
- Agent / 02
It is resolved. Please wait.
- Customer / 03
You said that last time.
- Repetition
- False resolution
- Escalation
What Nyeti returns
Five dimensions around every message
Nyeti reads the language, not the person. Treat each result as workflow input, not a diagnosis.
- 01
Emotional tone
The emotional direction of the message.
- 02
Mental effort
Signs the sender may need simpler guidance.
- 03
Urgency
How quickly the message may need attention.
- 04
Engagement
Signs of focus or disengagement.
- 05
Communication style
Formality and directness cues.
Where it helps
Use signals where decisions happen
Prioritise, guide, or hand off. Your team stays in control.
- 01
Prioritise for review
Flag messages that signal frustration or high urgency.
- 02
Guide the response
Pass posture, tone, and action guidance downstream.
- 03
Keep people in control
Combine signals with business rules and human review.
Where to start
Start where context changes the next step
Choose one repeatable message workflow and evaluate whether the added signals improve handling.
01
Support operations
Add urgency and emotional tone to the rules that prioritise an existing support queue.
02
Customer success
With conversation context enabled, follow changes in effort, engagement, and urgency across customer conversations.
03
Financial services
Add behavioural signals to human-review rules in transfer, verification, or account-access workflows.
NYETI API
Behavioural signals for customer conversations
Nyeti is a behavioural-signal API. It reads each customer message, returns five signals and response guidance, and helps support teams see where a conversation is going. It is a text API for customer support, not a clinical tool.
Five signals per message
One API call reads a customer message for emotional tone, mental effort, urgency, engagement and communication style, with response guidance. Results come back in under 5 ms.
Conversation analytics
Send turns with a conversation ID and Nyeti tracks the exchange: repeated requests, escalation, and resolutions that did not land.
DATA HANDLING
What happens to the text you send
No third-party AI providers
Results come back in under 5 ms. Your text is not sent to an outside AI provider.
Redaction on by default
Email addresses, phone numbers and card-number patterns are masked before text is stored for conversation history.
Scoped keys
Every call uses a workspace API key. Keys can be limited or revoked by workspace admins at any time.
Built for support text
Nyeti is a customer-support tool, not a clinical one. Please do not send health records through it.
AVAILABILITY
An API with plans, in closed beta
Approved teams get a workspace with scoped API keys, usage tracking and a plan that sets included monthly usage. Workspace admins can set limits per key. Applications are reviewed before access is offered.
Early access
See what Nyeti adds to one workflow.
Apply with one message workflow. Approved teams can evaluate the beta API before wider rollout.