feat(contact-center): Implementing get suggested response feature within CC SDK by Kesari3008 · Pull Request #4952 · webex/webex-js-sdk

Kesari3008 · 2026-05-05T20:32:19Z

COMPLETES #< INSERT LINK TO ISSUE >

This pull request addresses

Implementation for Get Suggestions feature

by making the following changes

Added new events to be handled under real time websocket connection listener callbakc
Added new method getSuggestedResponse exposed to users via cc.ApiAiAssistant object which invokes sendEvent() based on the received params.
Modified sendEvent to accommodate parameters required for GET_SUGGESTIONS and ADD_SUGGESTIONS_EXTRA_CONTEXT events and send correct pyaload so that AI assistant can start listening on conversation and give suggestions.
Sample app changes to add AI Assistant section.
Updated UT for the changes done.

Change Type

Bug fix (non-breaking change which fixes an issue)
New feature (non-breaking change which adds functionality)
Breaking change (fix or feature that would cause existing functionality to change)
Documentation update
Tooling change
Internal code refactor

The following scenarios were tested

https://app.vidcast.io/share/dc286450-408f-49d8-99b8-cf8b983db293?playerMode=vidcast

The GAI Coding Policy And Copyright Annotation Best Practices

I certified that

I have read and followed contributing guidelines
I discussed changes with code owners prior to submitting this pull request
I have not skipped any automated checks
All existing and new tests passed
I have updated the documentation accordingly

Make sure to have followed the contributing guidelines before submitting.

…responses feature

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f171b6631e

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector · 2026-05-07T08:09:34Z

+    const eventName = context
+      ? AIAssistantEventName.ADD_SUGGESTIONS_EXTRA_CONTEXT
+      : AIAssistantEventName.GET_SUGGESTIONS;


Derive suggestion event from trimmed context

getSuggestedResponse() chooses ADD_SUGGESTIONS_EXTRA_CONTEXT based on the raw context value, but sends context?.trim() to the backend. If callers pass whitespace-only input (common from text fields), this emits the “add extra context” event with an empty context payload, which can be rejected or ignored by the AI service instead of issuing a normal GET_SUGGESTIONS request.

Useful? React with 👍 / 👎.

chatgpt-codex-connector · 2026-05-07T08:09:34Z

+        case CC_EVENTS.POST_CALL_SUMMARY:
+        case CC_EVENTS.MID_CALL_SUMMARY:
+          break;


Emit summary RTD events instead of dropping them

This switch now explicitly swallows POST_CALL_SUMMARY and MID_CALL_SUMMARY RTD messages. Before this change, RTD payload types were re-emitted to the task, so these events would reach SDK consumers; after this commit they are silently discarded, which prevents any listener from receiving summary payloads even though these event constants were added to CC_EVENTS in the same change.

Useful? React with 👍 / 👎.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 48f08b217c

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector · 2026-05-07T16:51:49Z

+        case CC_EVENTS.SUGGESTED_RESPONSE_ACKNOWLEDGE:
+          // TODO: Handling this event
+          break;


Emit suggested-response acknowledge events to tasks

CC_EVENTS.SUGGESTED_RESPONSE_ACKNOWLEDGE was added as a task event, but this switch currently swallows it instead of re-emitting on the task. In environments where clients wait for an acknowledge before updating UI/request state, those listeners will never fire, so consumers cannot observe the backend acknowledgment despite the event being part of the exported event surface.

Useful? React with 👍 / 👎.

Shreyas281299

Mostly looks good. Just 1 think that might be a privacy issue.

Shreyas281299 · 2026-05-19T07:10:11Z

Do we need this ai-docs? These 4 files dont seem to fit with our ai-docs structure,

Shreyas281299 · 2026-05-19T07:19:50Z

+          break;
+        case CC_EVENTS.POST_CALL_SUMMARY:
+        case CC_EVENTS.MID_CALL_SUMMARY:
+          break;


Are these todos? If yes can we put some jiras for these todos

There are no JIRAs for these I think, I just added them because they will be implemented in future so I just added their case. I will remove the TODO comment

Shreyas281299 · 2026-05-19T07:21:43Z

+
+    return 'en';
+  }
+


I dont think we should be accessing global like this. Ideally we should have a default and then we can ask customers to set a language using a public method, rather than accessing their browser window. Might get flagged as a privacy issue

Also, this can be moved out to a util file. We should keep the core logic in this file

Agreed. I missed it's implementation, fixed it now

rarajes2

Based on the data processing complexity handled in the samples app, should we consider handling that inside the core sdk logic itself and provide a processed data to the end application for easy consumption ?
Please see if need to clean up the newly added ai-docs files
Also, the JIRA link is missing, please update the PR description

rarajes2 · 2026-05-21T05:00:00Z

 const transcriptEntries = [];
 const MAX_TRANSCRIPT_LINES = 200;
+const MAX_AI_ASSISTANT_ENTRIES = 50;
+const registeredTaskListeners = new WeakSet();


Q. Why are we using WeakSet and not regular Set ?

rarajes2 · 2026-05-21T05:00:39Z


 const transcriptEntries = [];
 const MAX_TRANSCRIPT_LINES = 200;
+const MAX_AI_ASSISTANT_ENTRIES = 50;


Q. Is this value set only for the samples page or this is the recommended value ?

rarajes2 · 2026-05-21T05:30:55Z

+
+function trimAiAssistantEntries(state) {
+  if (state.entries.length > MAX_AI_ASSISTANT_ENTRIES) {
+    state.entries.splice(0, state.entries.length - MAX_AI_ASSISTANT_ENTRIES);


Are we not keeping this data paginated ? Or is this because this info has to be short lived and agent does not have to scroll through to get the whole transcript at any time

rarajes2 · 2026-05-21T05:33:07Z

+  }
+}
+
+function getAdaptiveCardTextLines(node, lines = []) {


Should we add js-doc with info on the structure of data it receives for node and lines ? It will be easy to understand the implementation with that since it's a recursive function

rarajes2 · 2026-05-21T05:43:59Z

    appendRealtimeTranscript(payload);
  });

+  task.on('SUGGESTED_RESPONSE', (payload) => {


Should we keep the event name consistent with other casing ?

For AI assistant features, they are independent features so they are realy tied 1:1 with task. All the task related event are received in terms of AgentContact or similar but these events are handled in different manner and their payload also differs quite a bit. These events are also received on different channels. Due to these differences, we are keeping different naming convention for these events

rarajes2 · 2026-05-21T05:47:52Z

+          task.emit(payload.type, payload.data);
+          break;
+        case CC_EVENTS.SUGGESTED_RESPONSE_ACKNOWLEDGE:
+          // TODO: Handling this event


Do we have to address this in this PR itself or planning for future? We can add JIRA ticket link if needed to be handled later. We should add a log for the cases where it's silent

There is no Jira link for this. I added listening for this event just because this gets used to indicate whether AI assistant is in thinking or listening mode. But that's somehting we can keep internal to SDK hence in terms of implementation of this event, at present nothign is needed from the payload of this event but m not sure about future so just added the case for future refernce

Removed these events for now

rarajes2 · 2026-05-21T05:51:07Z

+
+    return 'en';
+  }
+


Also, this can be moved out to a util file. We should keep the core logic in this file

rarajes2 · 2026-05-21T05:54:33Z

+    LoggerProxy.info('Requesting suggested response', {
+      module: CC_FILE,
+      method: METHODS.GET_SUGGESTED_RESPONSE,
+      interactionId,
+    });


We can also add debug log for success and error log in the catch block. loggerContext can be reused as separate variable

rarajes2 · 2026-05-21T05:56:24Z

+        {
+          agentId,
+          orgId,
+          interactionId,
+          eventName,
+          trackingId,
+          context,
+        },


This info is repeated in success and failed cases. We can create a const with common fields

rarajes2 · 2026-05-21T05:57:32Z

+          if (
+            this.agentConfig.aiFeature?.realtimeTranscripts?.enable ||
+            this.agentConfig.aiFeature?.suggestedResponses?.enable
+          ) {


Should we update the associated TODO comment ?

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 5efa0fc964

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

chatgpt-codex-connector · 2026-05-22T13:44:29Z

+   * @public
+   */
+  public async getSuggestedResponse(params: SuggestedResponseParams): Promise<any> {
+    const {agentId, interactionId, context, languageCode} = params;


Default suggested-response language to English

getSuggestedResponse() forwards languageCode as-is, so when callers omit it the payload sends languageCode: undefined instead of the documented default ('en'). This changes behavior for the common path (no explicit language) and can cause backend routing/locale selection to be inconsistent or rejected if the service expects a concrete code. Add a local default (for example during destructuring) before calling sendEvent.

Useful? React with 👍 / 👎.

feat(contact-center): initial changes for implementing get siggested …

f171b66

…responses feature

Kesari3008 added the validated If the pull request is validated for automation. label May 6, 2026

Kesari3008 marked this pull request as ready for review May 7, 2026 08:04

Kesari3008 requested a review from a team as a code owner May 7, 2026 08:04

chatgpt-codex-connector Bot reviewed May 7, 2026

View reviewed changes

feat(contact-center): sample app changes

48f08b2

chatgpt-codex-connector Bot reviewed May 7, 2026

View reviewed changes

Shreyas281299 requested changes May 19, 2026

View reviewed changes

Kesari3008 changed the title ~~feat(contact-center): initial changes for implementing get suggested response feature~~ feat(contact-center): Implementing get suggested response feature within CC SDK May 20, 2026

feat(contact-center): remove unnecessary ai-docs

2a75395

rarajes2 reviewed May 21, 2026

View reviewed changes

feat(contact-center): review comments

5efa0fc

chatgpt-codex-connector Bot reviewed May 22, 2026

View reviewed changes

Conversation

Kesari3008 commented May 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

COMPLETES #< INSERT LINK TO ISSUE >

This pull request addresses

by making the following changes

Change Type

The following scenarios were tested

The GAI Coding Policy And Copyright Annotation Best Practices

I certified that

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot May 7, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector Bot May 7, 2026

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot May 7, 2026

Choose a reason for hiding this comment

Uh oh!

Shreyas281299 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rarajes2 left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector Bot left a comment

Kesari3008 commented May 5, 2026 •

edited

Loading

rarajes2 left a comment •

edited

Loading