AI Onboarding Roadmap and Project Plan

Your guide to enabling and exploring AI features in your environment

Step 1: Enable

Step 2: Verify

Step 3: Configure

Step 4: Test

Step 5: Deploy

Enable essential modules such as conversational analysis, translation, and FAQ management.

Use the admin dashboard to toggle features on and off.

Decide which features are most relevant to your organization’s immediate needs.

Ensure all necessary permissions are granted for users and admins.

Verify integration points for external systems (e.g., SharePoint or Azure Cognitive Services).

Check the status of each feature using the dashboard indicators.

Ensure that the system reflects “active” or “operational” for enabled features.

Use built-in test tools to verify basic operations like keyword extraction or chat summarization.

Validate key integrations (e.g., Azure OpenAI services and storage).

Ensure external services, such as Azure Translation, are connected and functional

Resolve any connectivity errors or incomplete configurations

Configure prompts, workflows, and pipelines to align with business objectives.

Set parameters for advanced features like cognitive search and summarization.

Use the admin console to assign roles and permissions for accessing AI tools.

Grant specific permissions to admins, agents, and end-users.

Link external data like SharePoint, FAQ files, and document repositories.

Validate indexing of external content to ensure availability during AI interactions.

Use a sandbox environment to preview AI behavior, such as response generation and chat workflows.

Test translations, FAQ lookups, and chat proxies in a controlled setting.

Share preview functionality with a small team of testers.

Collect feedback on accuracy, helpfulness, and ease of use.

Ensure responses align with business context and data.

Test against edge cases to evaluate AI reliability.

Roll out AI features to all employees and applicable guest interfaces.

Ensure adequate training and documentation are available for end-users.

Track usage metrics and initial feedback to identify areas needing improvement.

Set up automatic reporting to monitor adoption and effectiveness.

Provide a help desk for AI-related queries and troubleshooting.

Schedule regular updates and reconfiguration sessions based on evolving needs.

For a detailed step-by-step worksheet to keep track of tasks and progress for setting up your AI features, we also are providing this Onboarding Plan as a simple Microsoft Excel document:

Spreadsheet: Chime V5 AI Enablement Onboarding Plan

Basic Tenant AI Enablement and Configuration

Enable Chime V5 Core AI Features (Admin):

The first thing that you will want to do is to enable AI features for your tenant. To do this, as a tenant admin navigate from your tenant admin menu to the Configuration > Features. From the features page search for those from the list below and enable them.

ChimeV5.AI.AzureOpenAI

ChimeV5.AI.ConversationalAnalysis
ChimeV5.AI.PromptLibrary
ChimeV5.AI.Translation
ChimeV5.Pipeline.GptChat
ChimeV5.Pipeline.SummarizeConversation

ChimeV5.Pipeline.KeyPhraseExtraction

ChimeV5.AI.AgentChatAssist
ChimeV5.CognitiveServices
ChimeV5.FAQ
ChimeV5.FaqFiles
ChimeV5.SharepointFaqImporter

AI Features

View how our collection of AI features look within the Chime V5 UI.

FAQs: Create, Edit, and Manage FAQs

Within our manager UI users are able to create and manage FAQs in markdown format. These FAQs can be assigned tags, language, and can be indexed for AI searches.

FAQ Files: Upload Files as FAQs

Within our manager UI users can upload their existing files of various formats as FAQs. Like the markdown FAQs these FAQ Files can be assigned tags, language, and can be indexed for AI searches.

SharePoint FAQ Importer: Import SharePoint files as FAQs

Within our admin UI users can import SharePoint sites and pages as FAQs. These imported FAQs version and status can be viewed from the Import FAQs page and can be updated to the latest version in the Edit FAQ UI. Like normal FAQs these can be assigned tags, language, and can be indexed for AI searches.

Cognitive Search Indexing: Index your FAQs with Azure OpenAI Cognitive Search Services

In the manager UI in the FAQ creation and editor UI users can toggle on indexing for the content. Upon toggling ‘Should Be Indexed?’ on and publishing the file will be prompted to be saved to the blob storage and index of the Azure OpenAI Cognitive Search Service. This allows the content to be referenced in any AI generated responses.

Cognitive Search Indexing: View your file indexing status (Admin > Chime Settings > Azure Cognitive Search > Search Status Index > Content Items)

In the admin UI under Azure Cognitive Search > Search Status Index users can view indexing status for all FAQ files as well as specific FAQ Content Items. Under the Content Items tab users can view if an FAQ has been uploaded to blob storage, added to the KB, and indexing status.

Cognitive Search: Test AI Cognitive Search Service

In the manager UI users are able to simulate an AI FAQ Search query and preview the result set that a user would get back and the score based on how well the document content matches the query. This allows managers to streamline the process of confirming an FAQ will display in an AI search without running a pipeline.

Azure Key Phrase Extraction & FAQ Lookup: Extract Keywords from guest questions within a chat pipeline and search with FAQ Lookup

In the a pipeline the Azure Key Phrase Extraction stage can be added to evaluate a guest question and pull out keywords, entities, and language. By getting these metadata values, when navigated to the FAQ Lookup stage the user will receive a better result set of FAQs based on the keywords.

AI Assist: Generate AI messages to help Agents that reference indexed FAQ content

Within the agent web client UI under the AI Assist tab, agents are able to generate AI responses based on the chat conversation and guest query. These generated AI responses can be configured in the admin UI Chime Settings > Agent AI Chat Assist to edit the AI prompt, enable responses integrated with cognitive search that reference FAQs, and pull from that content as well as the number of suggested responses.

GPT Chat: Chat with conversational AI chat proxy and receive responses based on indexed FAQ content

In a pipeline the Chat GPT chatbot stage can be added for the guest user to proxy to an AI chat. Within the pipeline editor the various items about the chatbot can be edited including the AI Prompt, annotation formatting, number of referenced FAQs, and whether contextual snippets of the FAQ citations is turned on.

AI Summary: Summarize your chat both automatically in the chat pipeline and manually in chat session details

Within the session details UI as well as the pipeline stage, AI is able to auto summarize chat session conversations with an easy to process ‘Issue’, ‘Resolution’, ‘Summary’, and ‘Follow-up Tasks’ format. The pipeline stage can be added directly for the summary to be automatically populated or an agent or manager user could open the session details and manually load the AI Summary.

AI Reports

View our collection of AI integration reports. These reports are included as our set of prepopulated reports for user tenants with Premium and Enterprise service tiers that have AI services enabled.