The refreshed translation experience for real-time translation includes out-of-the-box translation services. New default settings for real-time translation include:

• Experience profiles that allow custom and default primary languages for incoming new messages.
• Configuration settings at the channel level rather than global, so each channel can have its own language settings.
• An out-of-the-box web resource that gives representatives a default real-time translation for their conversations from Microsoft.

New open configurations enable translation during conversations. Representatives can:

• Turn the service on or off for each conversation.
• Manually change the language for sending and receiving messages.
• Hide or view the original and translated versions of each text message separately to avoid confusion.
• Retry translations if there are issues.

Administrators can enable auto-promotion of intents at the line of business or intent group level by selecting semi-autonomous or fully autonomous promotion mode. Some of the key features are:

• Increase automation: Reduce the need for human intervention in intent promotion.
• Improve accuracy and trust: Make sure that only high-confidence, low-risk intents are promoted automatically by using robust validation mechanisms to prevent erroneous promotions.
• Enhance administrator control and flexibility: Allow administrators to configure self-promotion settings at various levels and provide override mechanisms.

Enablement and scope

• Available in Dynamics 365 Contact Center. Customer Intent Agent is part of the autonomous service agents suite and uses a consumption-based billing model on top of your base license for Dynamics 365 Contact Center.
• When the administrator turns on the feature in the Copilot Service admin center, Customer Intent Agent analyzes your existing support conversations and cases to automatically build an intent library.
• After the agent discovers intents, you can create a Customer Intent Agent in Copilot Studio. It uses your intent library to ask follow‑up questions, provide knowledge articles, or take actions to resolve customer issues end to end.

Enhancements

• Intent discovery customization: Customize intent discovery by supplying sample utterances, specifying key entities, and defining inclusion criteria to make sure intents align with your business needs.
• Hierarchical playbooks and orchestration: Tailor agent behavior for each intent by using natural language playbooks with no coding required. These playbooks support nested sub‑intents and conditional logic for complex workflows.
• Expanded channel and escalation support: Support both digital and voice channels through Azure Communication Services. The agent automatically escalates to the Case Management Agent for structured email follow‑up when needed.
• Action support via custom connectors: Run support actions, such as triggering refunds, updating order statuses, querying inventory systems, and automating non-Microsoft workflows through configurable connectors.
• Quality and monitoring enhancements: Real‑time dashboards highlight KPI deviations and provide intent‑level metrics. These features enable targeted adjustments and continuous optimization of agent performance.

• Intent determination: Use generative AI to recognize customer intent, a natural language description of customer problem, and intent groups—logical clusters of similar intents that represent the broad customer pain points and business expertise needed to address these pain points.
• Customizable user groups mapped to intent groups: Manually curate user groups to create clusters of support representatives with similar capabilities and expertise. Decorate groups based on attributes such as language, region, and other business criteria. Administrators can map user groups to intent groups and configure certain matching conditions if required.
• Intelligent assignment: Match customer inquiries to the most appropriate support representative in the user group based on real-time attributes like capacity and presence.
• Integration and flexibility: Easily integrate with existing customer service platform and provide flexibility to adjust routing parameters, adapting to changing business needs and customer demands.
• Supported channels: Live chat and persistent chat; case (Incident entity in record channel).

In Dynamics 365 Contact Center, administrators can enable Copilot to predict and generate resolution notes from case data in non-Microsoft CRM solutions like Salesforce and ServiceNow. Service representatives can choose to accept or reject these suggestions.

The notes follow the standard format of "Cause," "Symptom," and "Resolution." Copilot automatically generates them based on case history, emails, notes, and other important fields. These resolution notes address scenarios where a simple note needs to be added before a service representative closes a case.

Support teams can use AI-guided assistance to generate resolution notes directly with Copilot that's embedded within their own CRM solution. Agents get AI-powered suggestions to resolve incidents effectively and enjoy the following benefits:

• Save time: Automatically generated resolution notes reduce manual effort and eliminate repetitive tasks.
• Enhance accuracy: Copilot ensures that vital details aren't missed when creating resolution notes.

Copilot suggests resolution notes based on the incident context, helping capture the cause, impact, and resolution of the incident. The agent can select and edit these notes to finalize the resolution notes before submitting them.

When you use Copilot within your CRM solution, it generates prompts based on the context of the active case or conversation. Copilot creates prompts for the first question only and not for follow-up questions.

• For a case: Copilot considers the title, description, notes, and linked emails within the case in the CRM solution.
• For a conversation: Copilot uses the intent in the first three messages to create a prompt, then refreshes the prompts for every third message. Users can select a prompt to generate a response.

For organizations that use Copilot embedded in non-Microsoft CRM solutions, admins can now enable plugin support within the Copilot embedded widget. This support enables advanced functionality by connecting plugins that provide extra data inputs. Admins can configure these plugins and define context keys in the admin experience. Copilot uses these keys to deliver context-aware responses that align with the host CRM solution.

Customer feedback survey agent using Copilot Studio is a robust solution that simplifies and enhances the collection and analysis of customer feedback across voice, chat, email, and digital channels. This feature uses generative AI to tailor questions and interpret responses contextually. It includes the following key capabilities:

• Enhanced and optimized survey runtime experience for performance and usability.

• Integration with live chat workflows, making sure that the entire feedback experience remains within the same chat window—eliminating the need to redirect users.

• Configuration and enabling of the survey agents in Copilot Service admin center. You must publish the survey agents in Copilot Studio before adding them to workstreams.

• Customization of the surveys in Copilot Studio visual editor, where you define questions and branching logic.

The survey becomes a part of the runtime behavior. The system automatically triggers the configured survey for the customers at the end of the conversation.

Support agents can use Copilot within the embedded widget to summarize cases from their host CRM solution. Agents can do the following tasks:

• Summarize cases using attributes such as case title, description, product, priority, and key case information like email threads and case notes.
• Access the summaries in Copilot.
• Give thumbs-up or thumbs-down feedback to rate the quality of the generated summary and help improve Copilot generated summaries.
• Include verbatim feedback when giving a thumbs-down to share key information they want Copilot to include in the summary.

When agents select summarize case, Copilot automatically picks up context from the open case in the background. If agents invoke case summary without any open case in the background, Copilot prompts them to enter a case number to generate a case summary.

Copilot in Dynamics 365 Contact Center helps agents draft emails to send to customers by using email context from their CRM instance. Agents can use Copilot to:

• Draft email responses to customer issues and questions with Copilot in the Dynamics 365 Contact Center embedded and standalone apps.
• Use organizational knowledge from the CRM instance, and add details about the case related to the email with the email thread.
• Copy email responses from Copilot and paste them directly into responses to customers, with the option to review and edit them before sending the email.
• Give thumbs-up or thumbs-down feedback to Copilot-drafted emails, or provide verbatim feedback regarding the email quality to help improve Copilot responses.
• Use the out-of-the-box connectivity to Salesforce and ServiceNow systems to retrieve knowledge and then prepare emails.
• Use the context retention feature to switch among sessions and pick up the email draft where they paused.

Support agents can use Copilot to generate a summary:

• Of the entire conversation up to the point when they select the generate summary option.
• While the conversation is ongoing.
• When they wrap up the conversation.
• When they transfer the conversation to another agent.

The key features are as follows:

• Conversation Summary takes into account the conversation transcript.
• Agents can generate summaries for message and transcript-enabled voice conversations.
• Agents can give thumbs-up or thumbs-down feedback on the conversation.
• Support agents can provide verbatim feedback when giving a thumbs-down to vocalize key information they want Copilot to include in the summary. Supervisors and contact center managers can review the verbatims.

The timeline highlights are visible on the agent's timeline that you enable on any form within the Contact Center workspace.

• Agents get a shorthand capture of the key details and activities on the timeline, including notes, emails, posts, and any other timeline activities.
• Get an in-depth understanding of recent events on the timeline.
• View the highlights in a reverse chronological order.

With Copilot in Dynamics 365 Contact Center, agents can access their organization's knowledge, even if it's stored in third-party CRM solutions like Salesforce or ServiceNow. Copilot retrieves information from this knowledge base and helps answer questions about service-related issues. Agents can do the following tasks in the embedded widget or standalone app:

• Ask direct questions about customer issues and get answers from the knowledge base.
• Copy responses from Copilot and paste them directly into replies to customers, or review and edit them before sending the replies.
• Give thumbs-up or thumbs-down feedback for the Copilot responses, or provide verbatim feedback about the response quality to help improve Copilot responses.
• Use built-in connectivity to Salesforce and ServiceNow CRM solutions for knowledge retrieval.
• Context retention by Copilot in the embedded widget while a case is open in the background. Copilot resets the session if you navigate away and becomes active to where you left when you switch back to the session.
• In the embedded widget in Salesforce or ServiceNow, Copilot retains a global session context on any forms other than the case form or incident.

With Copilot-generated conversation summaries in GCC, agents can easily recap an ongoing chat or a transcribed voice conversation and efficiently resolve customer issues.

Messaging APIs provide a set of RESTful interfaces that you can use to control and manage customer conversations programmatically. With these APIs, you can start and end conversations and handle various customer interactions within a session.

A webhook subscription lets your services receive real-time updates about conversation events. For example, you can get notifications when an agent sends a message, when a system message is generated, or when an agent closes the conversation.

Key features of messaging APIs include:

• Starting a conversation with an authentication token and options for persistent conversations and context variables.
• Sending customer messages, including attachments and updates to context variables.
• Receiving agent or system messages through webhook subscriptions.
• Checking agent availability in a queue and retrieving queue position and estimated wait time.

Use messaging APIs when you need full control over the customer experience. They support extensive customization and work well in environments with strict network and data policies, such as closed network deployments.

In the complex journey of a contact center, customers often adopt or develop multiple systems that need to be in sync for optimal performance. Some scenarios include:

• Agents working on different channels in different systems – for example, contact centers using Dynamics 365 for messaging and another system for voice calls.
• Agents working on Microsoft Teams as well as Dynamics 365.
• Agents working across multiple Dynamics 365 environments.

The Presence API addresses these scenarios by ensuring synchronization across different systems.

Organizations can quickly onboard and use our communication capabilities (omnichannel) and unified routing-based work distribution with their existing CRM solutions like Salesforce. Administrators can connect their existing CRM solutions and use the channel capabilities in Dynamics 365 Contact Center with a one-time access setup to map the CRM solution data.

The key capabilities are:

• Identify customers and cases: Link unauthenticated users to CRM data for personalization, and match returning customers with existing tickets.
• Enriched service request: Enhance incoming service requests with additional context derived from the interactions, such as sentiment, and with existing data in the CRM solution, such as region of the customer.
• Route contextually: Route the service requests to appropriate queues and agents through customizable rules and real-time decision making. You can author the rules based on the customer-profile data in the CRM solution, such as VIP flag, current and past cases of the customer, and journeys.

This feature enhances contextual data exchange between voice agents and external systems in Dynamics 365 Contact Center environments using Direct Routing telephony. It's especially useful for voice agent makers and system integrators who need advanced data-sharing capabilities.

Previously, the SIP UUI field—with its 128-character limit—was the only method available for passing call context. This update introduces SIP X header support, enabling more flexible and scalable context transfer.

The feature offers two key capabilities:

• Receiving SIP X-Headers: Voice agents can access incoming SIP X header values by defining variables that match the header names.
• Sending SIP X-Headers: During external phone number transfers, voice agents can select up to five variables to pass as SIP X headers.

Technical limitations include:

• A maximum of five SIP X headers per call (for both incoming calls and external transfers).
• Header names limited to 64 characters.
• Header value limited to 256 characters.

Administrators can set up selective hold in Copilot Studio by following these steps:

1. Go to the question node's properties under the voice section when designing voice agents in Copilot Studio.
2. Turn on the Selective Hold option in the properties section to activate this feature for the specific question node.
3. Upload promotional prompts or music files that play while the caller is in the selective hold state.
4. Specify words or phrases such as "I got it" or "Got it" that callers can use to exit the hold state. Until the caller speaks one of these hold words, the music and prompts continue uninterrupted.
5. Set a timeout limit for the selective hold state, after which the voice agent automatically prompts the caller. If the caller doesn't respond within the specified timeframe, the voice agent automatically prompts them.
6. Set up multiple reprompts to engage with callers who stay in a hold state for long periods.
7. Combine selective hold with other Copilot Studio features and triggers to create comprehensive interaction flows.

During a call, callers hear continuous music or promotional messages until they indicate readiness with a predefined exit phrase. This approach creates a professional waiting environment without awkward silences.

Protection of sensitive customer information is a critical requirement for industries with strict compliance and data privacy needs. The sensitive data redaction capability in Copilot Studio enhances security and privacy by allowing AI agent makers to define sensitive variables and protect information assigned to these variables throughout the voice interaction lifecycle.

This feature applies to the voice channel and provides the following capabilities:

• Copilot Studio controls: AI agent makers can mark specific variables as "sensitive." Once designated, the system automatically excludes these variables from:
  • Transcriptions
  • Logs
  • Recordings
  • Any other storage or visibility in the system

These protections apply across various AI Agent components, including question nodes and message nodes.

• Service-wide application: The system consistently enforces masking rules across all services that handle human-to-AI (agent) communication, including:
  • Real-time transcriptions
  • Diagnostics logs and database entries
  • Call recording services

By automatically excluding sensitive data from visibility and storage, this feature strengthens your compliance posture, reduces the risk of data leakage, and helps your organization build trust with customers.

For voice agents (IVR) and voice applications, constrained grammar-based recognition delivers better accuracy, performance, and cost than traditional semantic-based speech-to-text (STT) AI engines. The engine processes grammars that follow the Speech Recognition Grammar Specification (SRGS) standard, which are usually written in an XML-based format called GrXML. The engine converts spoken audio (utterances) into text that follows the grammar's rules.

Constrained speech recognition works best for:

• Recognizing constrained lists such as addresses, stock tickers, zip codes, and department names.
• Recognizing alphanumeric strings like tracking numbers, account numbers, and confirmation codes that have positional constraints. For example, the first two characters of a member ID start with "AN", "FD", or "NT". Another example of positional constraints is vehicle identification number.
• Alphanumeric or digit recognition with checksums or similar constraints. For example, credit card numbers with Luhn checksums.
• Directed dialog applications requiring specific words or phrases.

Voice Agent Makers must perform the following steps:

1. Author GrXML grammar files within a rich text editor of choice that adhere to the Speech Recognition Grammar Specification standard.
2. After development, upload the grammar into an Azure Storage account.
3. In Microsoft Copilot Studio, create an external entity and use the URL of the grammar as the entity title.

During runtime, the voice agent loads the grammar, performs the recognition, and returns the required result.

Multi Entity Identification allows agent designers to attach more than one entity for recognition onto a question node in Copilot Studio, providing additional flexibility for customer self-service and leading to higher containment rates. This capability improves customer self-service by recognizing one of several valid inputs provided by a user, such as an account number or partial social security number. It also enables AI agents to offer alternatives when users do not have specific information, avoiding immediate escalation and enabling fallback options like additional questions or logic branches.

During runtime, if multiple entities are attached to a question node, the agent attempts to match the user's input to one of them, whether by voice or text. If multiple entities are detected, the agent matches only the first entity in the list. Designers should ensure prompt wording and follow-up logic account for this behavior. If no match is found, the agent follows the configured re-prompt treatment in the question node.

On the Copilot Studio authoring canvas, designers must perform the following steps:

1. Define and create the entities to capture.
2. Create a question node and craft the prompt to the user.
3. On the question node:
   1. Under Identify, select One of Multiple Entities.
   2. Select + New Entity and then select from the Entity picker.
   3. Repeat for all applicable entities.

All responses are saved in a variable of type “record” that can be renamed by the author.

This recognition method identifies only one entity per user utterance. If a user input contains more than one valid entity, only the first entity listed in the configuration is recognized.

For example, a question node uses Multi Entity Identification, where the entities are Tracking Number and Delivery Address. A user says, “My package to be delivered to One Microsoft Way with the tracking number 7658641290753147”. The result is 7658641290753147, as the tracking number is the first entity in the list.

To access specific information from the record object, subvariables are automatically created based on the possible entities captured by the question node and displayed in dot notation, {recordName}.{entityName}. This is used to create conditional branches following the Multi Entity Identification step to route the information captured.

We recommend using the Is (not) Blank triage option in the conditional nodes to proceed logically based on what information is received by the user at question time.

The following is a list of known limitations for Multi Entity Identification:

• Dual-tone Multi Frequency (DTMF) input for entity identification is not supported in this modality. However, it may exist for the entity on non-Multi Entity Identification-based question nodes based on the type of entity created.
• Currently, a maximum of five entities can be within the recognition pool on any given question.
• No external entity support.

With Power Apps, enterprise customers can use Dataverse tables to set up dynamic key values for their Copilot applications. Examples of these key values include store locations, transfer numbers, broadcast messages, and lines of business. By managing these AI agent rules, business administrators can improve operational agility and quickly adjust the voice agents (IVR) flow to meet their business needs. Using Power Apps removes the need to directly change the agent’s logic, saving time on standard testing and deployment procedures.

Customers can connect Power Apps with Copilot Studio in a few simple steps:

1. Create and configure your Dataverse table in Power Apps.
2. Create an application in Power Apps to support administration of your data.
3. Set up an action in Copilot Studio to connect your Dataverse table to your copilot.

When you set up the customer's application, business administrators can directly add, update, and delete key data values. They can also publish these changes to the live Copilot environment without any hassle. The data associated with the Copilot is cached, so users experience no latency or disruptions.

Key features include:

• A new report, the Omnichannel Real-time bot dashboard, that enables better monitoring of IVR applications in real time. Key capabilities of this feature include reporting of the following bot conversations:

  • Active
  • Closed conversations
  • Bot conversation duration
  • Failed bot conversations
  • Conversation escalations in real time
  • Ability to view call volumes by DNIS and the language of the bot after the conversation ends by IVR or chat bot.

  To enable the Omnichannel Real-time bot dashboard in your environment, contact Microsoft Support. For more information, see Get Help + Support.

  When you enable the feature flag, administrators must turn on the Customer Service Admin Center > Insights > Realtime Analytics > Add real-time analytics for bot checkbox to view the report.

• New metrics and filters are available for Omnichannel Historical Bot Dashboard. You can find these metrics and filters in the data model section of the Power BI dashboard. Only users with the Analytics Report Author role can edit the report. Call center supervisors can use the following tools for performance management, monitoring, troubleshooting, and optimization:

  • OutcomeReasons: Reasons why a conversation with the Copilot Studio bot ended. These reasons come from every session within the entire bot conversation. You can only see this value in historical dashboards through visual customization. You can add it as a matrix table to show the number of conversations based on their outcomes and outcome reasons. This table helps you identify human agent escalations by design, escalations due to a maximum number of attempts by the user, or if the user reached the IVR but abandoned the conversation.
  • DNIS: The specific phone number that users dial to reach the bot agent. It's a critical dimension for IVR reporting and operations. For example, when a user calls 800-555-0100, the system logs this number as the DNIS.
  • Number of failed calls: A metric that shows the number of conversations that failed to connect to an IVR after an attempt was made with a call failure treatment for a bot.
  • Language of the bot: Captures the last language used during the IVR interaction, which might differ from the default bot language in multilingual configurations.
  • Number of escalations to external phone number: An enhancement to the existing escalation metric that includes external transfers to Dynamics 365. This feature supports standalone IVR, which doesn't have Dynamics 365 agents.

Administrators can define the customer experience during unforeseen bot issues. Instead of abruptly ending the call with a busy tone, which is jarring for the caller, the application can gracefully manage the situation.

The application plays a courteous message such as, “We apologize, our application is currently experiencing technical difficulties. Please try your call again later,” or "We are facing some issues at the moment. Allow me to connect you with a representative who can assist,” guiding callers to the right course of action.

By default, this feature is activated when you create the voice channel, ensuring immediate readiness. Administrators can personalize the message to align with the organization's brand identity and customer engagement strategy. This customization can replace the standard message to reinforce the brand’s voice and enhance the overall customer experience.

Some of the options include:

• Play a custom prompt message
• Transfer the call to an external number
• Disconnect
• Play a custom prompt message and disconnect.
• Play a prompt message and transfer to a queue.
• Directly transfer the call to a queue.

In Copilot Studio, you can create a single agent that handles multiple languages. When you assign phone numbers in the voice channel in Dynamics 365 Contact Center, you can select a primary language for each phone number. This feature eliminates the need for multiple agents by linking languages to dialed numbers.

To handle calls originating in any language, configure a single multilingual bot in Copilot Studio and assign it to a common phone number for all users in your contact center. The bot logic can switch the bot language to the caller's preferred language based on the caller's language preference. This feature eliminates the need for multiple bots and improves CSAT by supporting multiple languages with one phone number.

The secure consult, transfer calls to PSTN numbers, and IVR feature enables customer service representatives to securely route calls to non-Microsoft systems while maintaining customer privacy and data protection. When representatives initiate consultations or transfers to external endpoints such as payment processors or verification systems, the platform automatically suspends recording and transcription to keep sensitive interactions secure.

Key capabilities include the following:

• Automatically pause recording and transcription when an agent consults or transfers the call to external phone numbers.
• Allow customers to send DTMF inputs to secure endpoints.

Admins can design rich media message templates that apply to both the live chat widget and WhatsApp. The following types are supported for each channel:

Live chat widget

• Forms: Messages with a series of questions of types such as multi-select, single-select, dropdown, single-line input, date picker, and multiline input.
• Suggested replies: Reply buttons.
• Custom JSON: Freeform cards that you create using Microsoft's adaptive card technology.

WhatsApp

• Suggested replies: A card with up to three reply buttons.

To get started, admins can easily create channel-aware templates in the Customer Service admin center app. The Customer Service workspace app automatically displays these templates, which are tailored to the specific channel. This functionality allows agents to seamlessly send the templates to customers. For features like live chat widget custom JSON, live chat widget forms, and WhatsApp suggested replies, the adaptive card JSON is ready for use in the adaptive card node of a Copilot agent. This setup streamlines the agent design by enabling rich media messages without the need to rebuild adaptive cards from scratch. When the customer responds and interacts with these messages, a response goes back to the agent and customer service representative.

This feature set empowers journey authors to design more intelligent and effective outbound engagement workflows by using Microsoft Teams phone numbers—whether provisioned via pay-as-you-go subscriptions, Operator Connect, or Direct Routing—for proactive customer outreach.

The feature enables seamless integration with enterprise telephony infrastructure while ensuring regulatory compliance. Journey authors can define dynamic caller ID rules based on customer attributes, allowing each call to present a trusted and contextually relevant identity, which improves answer rates and customer trust. To further enhance the experience, journeys can optionally reserve customer service representatives before call initiation, ensuring they're prepared and available, reducing wait times and increasing first-call resolution.

Enhancements to dial modes include a predictive dialer that automatically initiates calls based on agent availability and customer responsiveness. This dialer helps journey authors efficiently scale outreach across large customer segments. A preview timer gives service representatives a configurable window to review customer history before the call begins. This timer enables more personalized, informed, and effective conversations that drive better outcomes.

Dynamics 365 Contact Center automatically classifies proactive engagement calls into the following outcomes for all newly created proactive engagements and the existing proactive engagements:

• Busy
• Key SIT tones (reorder, vacant, intercept)
• Invalid number
• No Answer
• Answered by a human
• Answering machine reached

Supervisors can view insights built on this data from out-of-the-box real-time analytics or from the msdyn_proactivedelivery table for custom reports. Journey authors also see this outcome when they define the criteria for post call action in Customer Insights journeys.

Customer service representative data masking is a new opt-in feature for Dynamics 365 Contact Center.

When enabled, service representatives can select individual messages from the customer and redact them, so the customer's PII isn't stored in your repository. This feature complements the existing automatic data masking rules with both automatic and manual masking. You can be confident that your brand is protecting your customers' valuable information.

When an administrator turns on this feature in the Copilot Service admin center, administrators and journey authors can create proactive engagements within Dynamics 365 Contact Center that comply with the Telephone Consumer Protection Act.

Journey authors can specify if a proactive engagement is part of a commercial journey, like cross-selling or upselling. This setting ensures a customer service representative is available to talk with a connected customer in under two seconds after the customer accepts the call. Call data is available in Dynamics 365 Contact Center to use in your compliance reports.

In the Copilot Service admin center, you can now turn on the option to show previously closed conversations to your customers for persistent chat, chat SDK, and mobile messaging SDKs.

• The previous conversation appears before the latest conversation, an experience that customers intuitively understand from consumer messaging apps.
• Brands can opt in to this feature and choose how long to keep historic conversations.

Administrators can configure the following settings:

• Provision SMS mobile numbers through Azure Communication Services.
• Synchronize the SMS mobile numbers from Azure Communication Services on the Phone numbers page in Copilot Service admin center.
• Configure the Azure Communication Services-based SMS mobile numbers in workstreams to send to and receive messages from customers reliably.

The Native SDKs are native iOS and Android SDKs that enable applications to start, manage, and retrieve customer communications for Dynamics 365 Contact Center. Both SDKs contain two main components:

• A ready-to-brand, fully featured messaging user interface that developers can integrate into their existing application. The included interface is based on the live chat widget so brands can use the same features and customizations as their web program.
• An optional communication layer, containing all the messaging functions necessary to host a conversation that developers can use directly to support a new or existing fully custom messaging interface. Brands that choose to use the included interface don't need to interact with this layer.

You can get the SDKs from public repositories, just like the existing Chat SDK for web. To enable communications through the mobile SDK, an admin creates a new chat workstream in the admin center and references the workstream's ID in the code for the mobile application. We recommend push notifications, for which the admin can also configure Google Firebase and Apple Push Notification accounts in the admin center.

With Copilot Studio, administrators can create custom Natural Language Understanding models to predict the intent for incoming calls and extract critical information. This integration enables developers to upload complex, organization-specific NLU models to ensure the caller's intent is interpreted accurately, reflecting the unique jargon and vocabulary of the business.

The solution enables developers to design a tailored ontology for their domain by defining intents, entities, and the relationships between them through annotated data. Developers can build training sets with annotated samples based on this ontology and use them to train and test their language models.

The solution streamlines the import of large volumes of training data to minimize training time and accelerate the development process. Customers with a Dynamics 365 license get custom NLU support at no extra cost.

After you link your Copilot agent with Dynamics 365 Contact Center, select Microsoft Custom NLU Plus from the Language Understanding tab under Settings in Copilot Studio. Then, train your Copilot agent using the data you provide to the custom NLU.

Many industries, including healthcare and finance, require strict compliance with regulations that protect sensitive customer information, such as payment details and protected health information. Contact centers and customer service operations must ensure that they don't record or transcribe this data while still maintaining seamless customer interactions.

The Pause and Resume Recording and Transcription API enables businesses to temporarily stop and restart call recordings and transcriptions. It provides a compliant and secure way to manage sensitive interactions. This API ensures adherence to industry regulations while preserving customer trust and operational efficiency.

The Pause and Resume Recording and Transcription API provides the following key capabilities:

• API-based controls: Programmatically pause and resume call recordings and transcriptions.
• Web link integration: Allow agents to manage recording and transcription settings directly within their workflows.
• Visual notifications: Provide real-time visual alerts to agents when recording or transcription is paused or resumed.
• Regulatory compliance: Support legal and industry requirements for handling sensitive data.
• Operational efficiency: Minimize compliance risks without disrupting customer service operations.

By leveraging this API, businesses can ensure secure and compliant interactions, enhance customer trust, and streamline agent workflows, all while mitigating legal and regulatory risks.

When you integrate your existing telephony infrastructure with Dynamics 365 Contact Center through Azure Communication Services direct routing, you can bulk import your existing phone numbers. The Copilot Service admin center generates a template based on the data you enter, such as carrier, country, and number type. You can then enter your numbers in the template and upload them in bulk.

After the upload completes, you can view the status of your updated file. When the file processes successfully, the numbers are ready to use and appear on the Phone Number page.

The web engagement orchestration has two main components:

• Single script tag for live chat widget: Use the new script tag that isn't tied to any workstream but is instead tied to your organization and line of business. You can use this tag on any portal where you want to support messaging.

• Web engagement rules: Use the new page in the admin center to configure the chat widget according to your organization's business rules.

  • Take actions:
    • Engage customers by displaying a chat button or proactively opening the widget.
    • Set the look and features for the widget.
    • Add custom context to the conversation.
  • Run rules conditionally:
    • According to where the customer is in your website.
    • Based on your operating hours.
    • Based on the customer attributes that you provide.
  • Publish with confidence: Rules that support versioning and publishing.

The embedded contact center experience now includes conversation details, sign-out, sessions management, pick mode, and voicemail capabilities. Service representatives can take the following actions in the embedded experience:

• Assign conversations to themselves using the pick mode, adding more work items to their workload, along with sessions management.
• See customer and conversation details when conversing with customers, enabling them to view the holistic picture in real time.
• See voicemail and other async channels.
• Sign out of the embedded app.

To boost agent productivity, administrators can configure the following capabilities for conversations that come through the asynchronous messaging channels. In the Dynamics 365 Contact Center admin center, under Productivity, use the Timeout rules option to enable the following actions based on business service-level agreements:

• Send an automatic reply to a customer to engage them again if they're unresponsive after a certain period.
• Close conversations automatically after the configured idle time that's indicated by no customer responses, so the representative's capacity is released to help other customers in the queue.
• Send follow-up messages to customers to remind them to connect back to make progress on their support conversations.
• Reduce average handling time for service representatives.

Key capabilities of this feature include:

• Administrators can enable the application to generate a summary of closed conversations and voicemails.
• On the enhanced Closed conversation page, customer service representatives see an AI-powered conversation summary and insights for closed voice conversations and voicemails.

The IVR agents in Dynamics 365 Contact Center now support the en-AU locale. You can configure these languages in Copilot Studio voice agents for telephony in the voice channel. The IVR agents offer capabilities such as text to speech, speech recognition, and transcription, and can handle questions in the specified languages.

Administrators can configure the following settings:

• Provision SMS short codes through Azure Communication Services.
• Synchronize the SMS short codes from Azure Communication Services on the Phone numbers page in the Dynamics 365 Contact Center admin center.
• Configure the Azure Communication Service SMS short codes in workstreams to send and receive messages reliably with customers.

Organizations can now access diagnostic telemetry for the full lifecycle of a conversation through Application Insights, so they can troubleshoot runtime issues more effectively. This end-to-end data empowers teams to quickly identify problems, apply mitigations, and maintain seamless contact center operations.

Telemetry spans voice, messaging, and record routing channels. The key conversation stages covered are initialization, self-service, routing, representative engagement, and conversation end.

To get started, export Power Platform data to Application Insights. After enabling the feature, telemetry from conversations begins flowing in. Administrators and supervisors can use built-in queries—or craft custom ones—to analyze each conversation stage by filtering with a conversation or correlation ID found in the conversation URL. For every telemetry item, you can pinpoint where failures occurred and understand the reasons why.

If you prefer not to build queries, the out-of-the-box dashboards offer an intuitive path to insights.

Consult and transfer calls to any Teams user in your organization, regardless of whether they have a phone number assigned to them. Key features include:

• Use Voice over Internet Protocol (VoIP) to contact Teams subject matter experts in your organization.
• Teams users don't need to have a phone number.