Email Recommendation

Grace Wannapongsai
Grace Wannapongsai
  • Updated

Feature overview

Email Recommendation is designed for users who need to respond to emails quickly and confidently, especially when crafting a reply requires extra time to review context, choose the right tone, and structure a response. By generating a suggested draft from the current email and visible thread history within the user’s workflow, the feature reduces manual effort and helps users get to a high-quality starting point faster. This improves response speed, supports more consistent communication, and keeps users focused without switching tools.

In practice, this is most valuable in common but time-consuming scenarios, such as following up on missing documents, responding to negative or frustrated emails, answering product or service questions, or re-engaging inactive contacts. The result is a faster, smoother reply process that saves time while still leaving the user fully in control of the final send.

Main Capabilities

How it works

  • Scans the selected email and visible thread history to understand the latest message and surrounding context. 

  • Generates a recommended draft reply based on the parsed email context. 

  • Allows the user to refine the draft using rewrite options such as Rephrase, Longer, and Shorter. 

  • Allows the user to adjust tone using options such as Neutral, Professional, and Casual. 

  • Supports additional refinement through an optional follow-up prompt.

  • Inserts the Recommendation directly into the draft reply for final review. 

  • Keeps the user in control to review, edit, and send the response manually.

Feature highlights  

  • Context-aware drafting: Uses the current email and visible thread history to generate a reply aligned to the latest message (see Limitations for Outlook mail layout considerations).  

  • Flexible refinement: Allows users to adjust the draft using rewrite actions and tone controls.

  • Grounded in the current conversation: Generates recommendations based on the selected email and visible thread history available at the time of drafting.

  • User-controlled workflow: Generates a draft for review, allowing the user to edit and send the response manually.

User Impact

Accessing Email Recommendation

First, open an email thread. Then, simply click the Email Assistant button located under Actions tab in Riva Insight panel.

Generating a Recommendation

Once opened, the feature analyzes the selected email and visible thread history to generate a Recommendation.

generated_reply_full.png

Editing a Recommendation

After a reply is generated, users can refine it using rewrite options, tone adjustments, or an additional prompt before inserting it into the email.

editing_reply.png

Inserting the generated reply

Once finalized, click the Insert button to insert the generated reply into the email draft for final review before sending.

email_inserted.png

Responding to short emails

When an email is brief, such as “Thanks”, Email Recommendation can help draft a concise response based on the available thread context. Users should review the suggestion carefully, as limited context may affect relevance.

reply_to_thanks.png

Responding to detailed follow-up emails

When an email includes multiple questions, requests, or next steps, Email Recommended Reply can help draft a structured response that addresses each point clearly and keeps the conversation moving forward.

detailed_follow_up.png

How Email Recommendation helps

  • Following up on pending requests
    When an email asks what documents are still needed, the feature can generate a reply that reminds them what is outstanding, answers the immediate question, and suggests next steps.

  • Responding to a negative or frustrated email
    When an email expresses dissatisfaction, the feature can draft a calm, empathetic response that acknowledges the concern and proposes a next step such as a call or review.

  • Answering product or service questions
    When an email asks about fees, services, or policies, the draft can provide a concise answer if the thread contains enough context. If not, user review is expected before sending.

  • Reconnecting with an inactive contact
    The feature can help create a warm, professional outreach email to reconnect with a new or dormant client and suggest a next step.  

Limitations

Currently not supported

Email Recommendation does not currently:

  • use CRM data as context

  • use other email threads as context

  • generate a separate follow-up email draft beyond the reply-drafting workflow

  • analyze attachments, files, photos, or other embedded media; recommendations are based on the email message body and visible thread history

Context availability

Context scope depends on Outlook mail layout settings.

  • When Outlook is set to Do not group messages, Email Recommendation can use the full visible email thread. 

    outlook_layout_setting_doNotGroup.png

  • When Outlook is set to Group messages by conversation, it uses only the currently selected message. Selecting a different message in grouped view refreshes the feature and requires Email Recommendation to be started again.

To learn more about changing this setting, refer to View email messages by conversation in Outlook

Where results may be weaker

Quality may be lower when:

  • the latest message is extremely short, such as “Thanks” or “Ok”

  • the thread is very long or heavily quoted

  • thread history is partial or truncated

  • repeated follow-up refinements cause the response to drift away from the actual email context

  • follow-up prompts are non-sensical or ambiguous 

Intended use

Email Recommendation is designed for email reply drafting only. It is not intended for unrelated or general-purpose requests.

Enablement

Interested in these features? Contact your Riva Account Representative or reach out to Riva Support Team.

Trust & Transparency

Best results come from clear context

Email Recommendation performs best when the latest email is clear and the visible thread history provides enough context. Results may be more generic or less relevant when the thread is truncated, messy, or lacks enough signal.

Data handling

Email Recommendation captures the current email content and visible thread history to generate a draft response. It does not store this information after processing.

AI-generated output may vary

Because the feature uses AI, generated drafts may differ from one attempt to another. The goal is to provide a relevant, high-quality starting point, even if the wording is not always the same.

User review remains essential

Email Recommendation is designed to assist with drafting, not send emails automatically. Users should always review the content for accuracy, tone, and appropriateness before sending.