Skip to navigation
PolyAcademyPolyAcademy level 2

Tutorial: Response control

Level 2 – Lesson 4 of 8 – Master response controls to regulate agent output and enforce tone.

Lesson
Level 2 · Intermediate · 15 min

Advanced voice settings regulate the agent’s own output. They run after the model generates text and before anything is spoken or sent. Useful for tone, safety, and flow – easy to misuse if patterns are too broad.

What response controls are

Response Controls are string- or regex-based rules that inspect the agent’s generated response.

Halt responses

Prevent specific phrases from being spoken

Trigger routing

Activate corrective or fallback behavior

Log patterns

Track unwanted patterns for analysis

Enforce tone

Maintain brand voice consistency

Response controls are not about what the user says – they are about what the agent says.

Which tool to use for which problem

ASR mishears user Wrong topic retrieved Agent says something it shouldn't Yes No, just monitor it Yes No Unwanted agent behavior What type of problem? Use ASR tools / Keyphrase Boosting Fix Managed Topic name or samplequestions Use Response Controls Should phrase NEVER appear? Halt – set Say phrases: TRUE Log – set Say phrases: FALSE Needs a fallback action? Attach a function Done

When to use response controls

Use response controls when you want to fix or prevent:

Examples:

  • “Let me check that for you…”
  • “Please hold while I look into this…”
  • If the issue is about recognition, use ASR tools
  • If the issue is about retrieval, fix the FAQs
  • If the issue is about how the agent speaks, use Response Controls

How response controls work

1

LLM generates response

The model creates text based on context and prompts

2

Response Controls scan output

Regex patterns are evaluated against the generated text

3

Action taken if matched

  • The response may be halted
  • A function may be triggered
  • The event may be logged for analytics

Controls are evaluated in real time and apply consistently across Chat and Call.

Check your understanding

  • Managed Topic sample questions
  • Keyphrase Boosting in Global ASR
  • A Response Control with Say phrases set to TRUE
  • A Transcript Correction rule

Answer: A Response Control with Say phrases set to TRUE

This is an agent output problem – the agent is generating filler phrasing. A Response Control with ‘Say phrases: TRUE’ will halt the unwanted phrase before it reaches the user.

Core fields

Each Response Control includes:

A short, descriptive identifier used for tracking and debugging.

Optional, but strongly recommended. Explains why the rule exists.

The exact pattern that will be matched against agent output.

Despite the name, this field controls halting, not allowing:

  • ON (Halt): matching output is blocked before it reaches the user
  • OFF (Log): matching output is allowed through but recorded for analytics

A fallback action to run when the phrase is detected.

Best practices

Start narrow

Broad patterns are hard to debug and can suppress valid responses.

.*sorry.*

\b(sorry for the inconvenience|apologies for the delay)\b

Target symptoms

Encode specific phrases you have observed, not abstract concepts.

“Prevent the agent from being verbose”

\b(let me explain how|to give you some background)\b

Prefer halting

If a phrase should never appear, halt it rather than replacing it.

Test thoroughly

Verify in both Chat and Call to catch unnatural truncation.

Common use cases

LLMs often add filler before answering or executing actions.

Example unwanted output: “Let me check that for you. Please hold…”

Suggested control:

  • ID: flow_redundancy_cutoff
  • Regex: \b(let me check|please hold|one moment while I)\b
  • Say phrases: TRUE

Result: The agent proceeds directly to the answer or action.

Logging-only controls

Not all controls need to halt output. You can use Response Controls purely for monitoring.

  • ID: hedge_language_monitor
  • Regex: \b(might|maybe|possibly)\b
  • Say phrases: FALSE

Tracks hedging frequency without disrupting users.

Testing and verification

After adding or updating a Response Control:

1

Test in Chat

  • Confirm the phrase no longer appears
  • Ensure the response still completes cleanly
2

Test in Call

  • Listen for unnatural truncation
  • Ensure turn-taking still feels natural
3

Review Conversation logs

  • Check whether the rule is triggering too often
  • Look for false positives

Common mistakes

Avoid these pitfalls:

  • Using overly broad regex patterns
  • Blocking phrases that appear in legitimate answers
  • Using Response Controls to fix retrieval problems
  • Forgetting to test in voice after chat validation

Check your understanding

  • What the user is allowed to say
  • How ASR processes audio input
  • The agent’s own generated output
  • Which FAQs can be triggered

Answer: The agent’s own generated output

Response Controls run after the LLM generates a response, before it’s spoken or sent. They regulate the agent’s own output – not what the user says or how speech is transcribed.

Pronunciations (separate feature)

Response Controls

Affect what is said

Pronunciations

Affect how it is spoken

Use Pronunciations when:

  • Names are mispronounced
  • Numbers need pacing
  • Domain terms require IPA or SSML guidance

Do not use Response Controls to solve pronunciation issues.

Response Controls are a surgical tool – the final layer of polish, not the first fix.

Try it yourself

1

Challenge: Write a response control

You notice the agent frequently says “Let me look into that for you” before answering – which creates an awkward pause in voice calls.

Write the full Response Control configuration:

  1. ID
  2. Description
  3. Regex
  4. Say phrases setting

Write the regex to match the specific phrase you observed, not a broad pattern. Start as narrow as possible – you can always widen it if needed.

  • ID: filler_preamble_cutoff
  • Description: Prevents the agent from saying “Let me look into that for you” before responding – this creates latency and sounds unnatural in voice.
  • Regex: \blet me (look into|check) that for you\b
  • Say phrases: TRUE

Check your understanding

  • Halt monitors the phrase; Log blocks it
  • Both do the same thing but at different latency costs
  • Halt blocks the response; Log lets it through while tracking it
  • Halt applies globally; Log applies per conversation

Answer: Halt blocks the response; Log lets it through while tracking it

If a phrase should never appear, Halt it – the response is blocked before reaching the user. Logging lets the phrase through while recording it, which is useful for monitoring but not for enforcement.

← Previous: Return values

Lesson 3 of 8

Next: Audio management →

Lesson 5 of 8