21 posts tagged with "AI"

This release brings new campaign management tools, additional KPI for realtime reports and voice related stability improvements.

Change log​

New functionality​

1. Voicemails are now displayed in the administration portal. Voicemails left by customers through the Voicemail node are now displayed in the administration portal. Supervisors or team leads can now see their details, playback voicemails and download them via the dedicated menu item. Each voicemail is displayed as a separate row in the list, even if there are multiple voicemails for the same conversation. Users may search for voicemails by conversation ID or customer phone number.

2. We added a new KPI for email-heavy businesses. You can now view, in the team reports, the total emails sent by your team as well as per agent in the agent view. This KPI is displayed at a glance, even if the agent card is expanded.
3. Automatically move agent to a selected break when entering lookup task mode. As a supervisor or team lead, you can now configure an option to automatically move agents to a break code you choose when they open their lookup task. This break code is not visible in the agent application as a selectable option and is automatically cancelled when the lookup task ends, without moving the agent to break. This is particularly useful if your agents perform back-office tasks and they should not be disturbed with an inbound task while they are in a lookup task.

4. Change predictive campaign dialing mode to preview. While operating a predictive campaign, you may decide that the campaign is not very efficient due to low agent count, or it already achieved its purpose, in which case, from the predictive settings, you may switch the campaign back to preview, without having to create another campaign. You can switch back to predictive at any time, but records already in progress are not affected by the change.
5. In some cases, the default music for customers waiting in the queue to be connected to an agent when using predictive dialing mode is not the best approach. For this, we introduced the option to disable it completely via the predictive settings, in which case, the customer will not hear anything until connected to an agent.

6. We introduced the option to enable queue timeout for predictive campaigns. When enabled, the timeout can be set from 0 seconds (a single attempt is made to connect the customer with an agent before abandoning the call) to 60 minutes. Customer calls terminated by a queue timeout will be considered a nuisance, the record state will become postponed and their dialing mode will be switched to preview.

Improvements and adjustments​

1. Fixed an issue where a predictive campaign record was being set to Rescheduled instead of Postponed when the call was flagged as a nuisance call.
2. Fixed an issue where the Language profile creation modal was not scrollable on small screens.
3. Fixed an issue where the CSV export for historical conversation reports was missing some fields.
4. Fixed an issue where announcements could not be stopped if there was a transfer happening while the announcement was playing.
5. Fixed a rare issue where an agent could receive two calls at once.
6. Further resilience improvements for outbound calls.

This release brings message status indicators for webchat, new campaign functionalities and additional agent management tools.

Change log​

New functionality​

1. We introduced the Read receipts functionality for webchat conversations. This allows the agent and the customer engaged in a conversation to see if their message has been sent, delivered, and read by the other party. Please note that this functionality is not available for other digital channels, such as Instagram, WhatsApp, or Facebook Messenger.

2. We introduced the option to Pause new campaign record processing to campaigns. This allows campaign administrators to ensure rescheduled records are processed and agents can meet their commitments to customers. When this feature is enabled, newly added or postponed records are not processed; only rescheduled records are handled.

3. A second campaign management tool we introduced in this release is the option to Suspend campaign records. Suspending a campaign record prevents it from being processed until unsuspended, allowing team leads to temporarily restrict records from being handled. This operation is available per record through campaign actions as well as in bulk through the selection mode.

4. The option to suspend campaign records is also available through the public API: you can suspend and unsuspend campaign records by their IDs, up to 100 per request.

5. BI developers need to track consultative call attempts accurately. We've simplified this by introducing a relationship ID in the OData calls table, allowing them to link consultative call attempts to their originating calls.
6. Agents may sometimes inadvertently cause issues with customer records. We introduced two controls for administrators to restrict access to certain functionalities. Administrators can now prevent agents from creating or editing customer records, helping keep customer data accurate. This is fully optional and configured per tenant.

Improvements and adjustments​

1. We've improved how campaign filters work. Previously, users could select only one campaign record state as a filter, but with the introduction of campaign record suspension and its associated filter, this became unnecessarily complicated. We've enhanced the user experience by transforming the status filter into a multi-select dropdown grouped by active and completed states. You can now apply multiple status filters combined with other filters, making it easier to find the records you need.
2. Emails received by the system are now marked as read. Previously, if the administrator checked the mailbox configured as an inbound email channel, emails processed by the system were not marked as read. Administrators had to check each email individually to verify processing. Emails processed by the system are now automatically marked as read in the mailbox.
3. We modified workflow-controlled campaigns to prevent system spam from very short calls. The maximum parallel call count is now guaranteed for calls longer than 30 seconds. For shorter calls, the system will not initiate new calls until the 30-second mark is reached.
4. Additional redundancy for voice services, rolled out gradually.
5. Fixed an issue where replying to a transferred email kept the original channel email address as the sender.
6. Fixed an issue where email conversations could be stuck without history updates or failed routing in situations involving faulty conversational bot configuration and handling.
7. Fixed an issue where a tenant having multiple Entra IDs configured could not create new channels under specific circumstances.
8. Fixed an issue where cloning a workflow did not clone its containers as well.
9. Fixed an issue where logout durations happening over the weekend were not tracked correctly.
10. Fixed an issue where unquoted display names resulted in the email address not being parsed correctly when pasting them into the email to field.
11. Fixed an issue where the agent workspace ID was missing from the OData feed for calls initiated through the public API.
12. Fixed an issue where a call freshly associated with a new contact record was not shown in the activity history.
13. Fixed an issue where the recording was not available for calls initiated through the public API.

This release brings a new public API, the option to add multiple email addresses in the TO field and additional improvements.

Change log​

New functionality​

1. We added a feature that enables campaign record status updates to be populated in the OData feed. It can be enabled for your tenant on request.

2. We published an API to cancel a campaign record by the record ID. It allows cancelling records one by one or in bulk, up to 100 records per request. See the developer's documentation for more details.
3. We added the option to use workflow variables in the Redirect To Phone node, enabling dynamic population of the phone number field.

4. We added the option to add multiple email addresses in the TO field of an email, up to 50. Additionally, we now support the option to paste a list of email addresses in the To/Cc/Bcc field. The user interface will automatically prompt the user to verify the email addresses when doing so.

5. Predictive calls are now automatically answered on the agent side. This enables the agent to save a couple of seconds and improve the overall predictive campaign performance.

Improvements and adjustments​

1. Fixed an issue where the email conversational bot's conversation history window contained previous email segments, modifying the email bot behavior.
2. Fixed an issue where users could not change the translation engine or remove the dictionary, when there was a dictionary assigned to a Language profile pair.
3. Fixed an issue where invalid email addresses could be pasted into To/Cc/Bcc without a warning.

This release brings new integration possibilities, bug fixes and the agent announcements feature.

Change log​

New functionality​

1. You may now specify multilingual voice conversational bot welcome messages. This feature was previously available for all the voice workflow nodes with the addition of the Workflow TTS and multilingual support. It is now added to the conversational bot's welcome message, removing the need to create a separate announcement to be able to handle multiple languages for the same message.

2. You may now categorize calls based on the live transcript instead of the recording transcript. This is particularly useful when post-call recordings are not available due to corporate policies or licensing restrictions, allowing categorization to happen based on the live transcript generated by a conversational bot, call translation or call transcription.

3. We renamed the Conversational API tools to Bot tools and added JavaScript support as a bot tool. Previously, integrators were able to configure an API integration for conversational bots to use. Now they can select between API and JavaScript, JavaScript allowing integrators to write their integration directly into the portal without needing to create, deploy and maintain a dedicated middleware.

4. We now allow integrators to request access for the external webpages they choose to use as task addons in the Agent workspace.

5. We added the option for team leads to configure automatic greetings when the agent answers an inbound call. The greeting will be automatically played, saving the agent the effort to read a greeting, or disclaimer, word by word.
6. We added the option for team leads to also configure on-demand announcements the agents can use based on the Agent Workspace the queue is using for an inbound call. These act similar to canned responses for chat or email templates for emails, allowing the agents to simply click an announcement instead of reading it themselves from a script. The agent can stop the announcement at any time.

7. We published a new public API to initiate and drop calls on behalf of an agent. Read more about it on our developers' website.

Improvements and adjustments​

1. Fixed an issue where the administration portal showed incorrect error messages when browsing through the campaign listing interface.
2. Fixed an issue where the customer identifier could display incorrectly in some cases.
3. Fixed an issue where the JavaScript node incorrectly chose a successful path when the code itself had a syntax error.
4. Fixed an issue where the Conversation API from our developer website could not be tried out on the website.
5. Fixed an issue where the browser would erroneously show a Leave popup warning when downloading an attachment over chat.
6. Fixed an issue that caused an incorrect tab to be activated when the agent refreshed the agent application.
7. Fixed an issue where there was an erroneous validation error on the Bot tools creation page.

This release brings new features, improvements, and fixes to make the system easier to use and more stable.

Change log​

New functionality​

1. Added portal-based testing of custom task wrap-up notes by work item ID, with validation and quick copy action.
2. Added a configurable delay for auto-answering inbound calls, including per-agent override and countdown with manual pickup option.
3. Added a tenant-level limit for parallel flow-controlled outbound calls with validation and start-blocking when limits are exceeded.
4. Added the ability for supervisors to assign agents to business unit–specific breaks based on shared business unit membership.
5. Added an option to disable automatic creation of customer records per channel for unidentified customers.

Improvements and adjustments​

1. Fixed an issue where agent chat messages were not delivered or visible when customers backgrounded the conversation on smartphones, causing missing messages upon return.
2. Fixed an issue where external web page links did not automatically open in a new tab upon conversation routing, despite being configured to do so.
3. Fixed an issue where real-time reports experienced delays in loading data after login or page refresh.
4. Fixed an issue where accounts could not be created when email addresses contained punctuation or non-English characters.
5. Fixed an issue where end time fields (EndDate and EndDateUTC) were missing in OData feed tables for agent states.
6. Fixed an issue where enabled auto-answer settings did not automatically present inbound calls to agents in voice configuration.
7. Fixed an issue where outbound dial CANCEL requests did not include the required X-Tenant-Id header.