Questions & Answers

Questions & Answers (Q&As) refer to a collection of commonly posed questions and their answers regarding a particular product, service, or topic. The Questions & Answers (Q&A) feature on DevRev works as an additional knowledge source for Turing bot to answer customer queries in conversations to provide efficient customer support.

Turing answering behavior

Unlike knowledge articles, Q&As are never quoted as sources to the end user. Q&As are preferred to answer customer queries over articles.

Q&As can be created either manually or automatically from your customer conversations by Turing.

Manual creation

In the following cases, you may choose to create Q&As manually:

• You can create Q&As manually for website FAQs that you prefer not to display as articles but are crucial for Turing to access to effectively respond to customer inquiries.

• You can incorporate internal knowledge base content into Q&As for internal documentation that's not intended for public search but is valuable for Turing to understand, enhancing response quality.

• You can include details about bugs or issues in Q&As that shouldn't be publicly searchable in your knowledge base and are not relevant to all customers but are essential for Turing to address specific customer questions.

Creating a Q&A

1. Go to Settings > Knowledge Base > Questions and Answers.
2. Click +Q&A in the top-right corner.
3. Fill in the Question, Answer, and select the relevant Part.
4. Set the appropriate Status and Access level. Set the status to External or Public if you want Turing to answer using it.
5. Confirm by clicking Create.

Automatic creation

Your customer conversations include a lot of knowledge, including some of the most accurate and current information. With the automated creation of Q&As, Turing learns from customer conversations in which a human answers a question that Turing previously could not answer.

flowchart LR
    A[Customer query] -->|Answer using published Q&As| C{Turing behavior}
    C -->|Satisfactory answer| D[Conversation resolved]
    C -->|No answer or talk to engineer| E[Creates Q&As] --> G[Improves Turing] --> C
    C -->|Similar Q&A found| F[No new Q&As created]

Enabling automatic creation of Q&As

To enable automatic creation, go to Settings > Knowledge Base > Question and Answers and click Preferences on the top right.

Enable Auto generate Q&As and click Save.

You need to be an admin to set preferences

How Turing learns from customer conversations

When a customer initiates a conversation seeking answers, Turing springs into action, drawing upon published Q&As and knowledge articles to provide a solution. If Turing falls short or the user prefers a more personalized touch, they opt to connect with a customer experience engineer and resolve the conversation. Turing doesn't just move on—it learns. Autonomously, it generates new Q&As based on the resolved conversation, marking them for review under the Review Needed status. A Q&A won't be created if it is similar to the existing Q&A to avoid duplication.

Reviewing and publishing Q&As

The conversation's owner receives a notification regarding these newly created Q&As. It's their chance to ensure accuracy before deciding whether to Publish them if needed or Archive if not.

Once approved and published, these Q&As become Turing's knowledge base, ready to tackle similar questions in the future. Each conversation can generate up to three unique Q&As, increasing Turing's knowledge base with every interaction.

Every Q&A is intricately linked to its originating conversation, preserving the context of its creation. Even if initially overlooked, support team members can revisit pending Q&As from Settings > Knowledge Base > Question and Answers filtering for those in the Review Needed state.

Searching for Q&As

Search for a previously created Q&A on the DevRev app by pressing cmd+K on Mac or Ctrl+K on Windows. Q&As appear within the KB section of the Search bar. You can search Q&As by entering keywords within a Question or Answer.

Components of Q&A

• Question: Define the customer query.
• Answer: Provide the corresponding solution or information.
• Part: Choose the product or service area the Q&A pertains to.
• Status: Determine the progress phase of the Q&A - Published, Draft, Review Needed, or Archived.
• Access Level: Determine who can view and access the Q&A.

Status definitions

• Published: These Q&As are live and actively used by the Turing bot to answer customer queries in conversations.
• Draft: Indicates that the Q&A is still being written or is a work in progress.
• Review needed: If the author requires validation, the QA can be marked for review, typically by a manager or designated reviewer.
• Archived: Denotes Q&As that have become redundant, incorrect, or obsolete.

Access permissions

• Private / Internal / Restricted: Not visible to external users, not used by the bot, and only searchable within the app.
• External / Public: Available for all, the Turing Bot employs these Q&As extensively, and they appear in search results both for app users and portal and PLuG visitors.