About Mogli AI
Mogli AI currently offers two powerful tools:
Conversation Summary: Enables you to quickly catch up on conversation history between your organization and your stakeholders
Message Recommendation: Keeps your conversations moving forward with intelligent response suggestions
Mogli AI has powerful capabilities - unlock it’s full potential by configuring metadata & prompts, creating great prompts, and utilizing reports and dashboards that get you the information you need when you need it.
Example Use Cases
🎓 Education
Conversation Summary | Message Recommendation |
|---|---|
Advisors supporting college-bound students can leverage Conversation Summary to efficiently track progress, allowing for more personalized guidance and focused support on crucial next steps. | Advisors can leverage the tool to provide timely and relevant guidance to students. By analyzing past interactions, the system suggests appropriate follow-up messages, ensuring students stay on track with their academic goals and receive personalized support. |
👍 Customer Service
Conversation Summary | Message Recommendation |
|---|---|
Customer support agents, handling a multitude of customer interactions, can utilize Conversation Summary to efficiently track each customer's issue and resolution journey. By quickly extracting key points and action items from support conversations, agents can provide more personalized and proactive assistance, ensuring customer satisfaction and timely resolution of their inquiries. | Support agents can utilize the tool to enhance customer interactions by receiving recommendations for the most effective next message. This enables agents to provide faster, more accurate responses, leading to increased customer satisfaction and quicker issue resolution. |
👾 Technical Support
Conversation Summary | Message Recommendation |
|---|---|
Technical support teams can use Conversation Summary to condense complex troubleshooting dialogues. This quickly identifies initial problems, diagnostic steps, and error messages, streamlining issue resolution and enabling efficient escalations. | Technical support teams can streamline troubleshooting by using the tool to suggest relevant diagnostic questions or solutions. This allows technicians to quickly guide users through complex issues, reducing resolution times and improving overall efficiency. |
🛒 Marketing
Conversation Summary | Message Recommendation |
|---|---|
Marketing teams can use Conversation Summary to analyze individual customer feedback and campaign engagement. Marketers can quickly qualify leads by extracting key phrases and sentiment. | Marketing teams can optimize lead nurturing and customer engagement by using the tool to recommend personalized messages based on previous interactions. This helps deliver targeted content and promotions, increasing conversion rates and strengthening customer relationships. |
💰 Sales
Conversation Summary | Message Recommendation |
|---|---|
Conversation Summary can ensure that sales professionals have consistent and timely follow-ups with leads by summarizing their interactions and providing useful information. | Sales teams can follow up with prospects using AI-recommended messages that are relevant and personalized to the conversation’s flow. |
🩺 Healthcare
Conversation Summary | Message Recommendation |
|---|---|
Conversation Summary improves healthcare by allowing providers to quickly reference previous patient visits, which leads to more personalized patient care. | Care team or admin staff can send appropriate reminders for appointments, educational opportunities, or billing matters while leaving out any private health information. |
Included In The License
The license covers installation and activation for both one sandbox and one production Salesforce org, along with the essential setup of the foundational AI model that powers these features. You also get flexible configuration options, allowing you to apply these tools to any Mogli-integrated object within your Salesforce environment. Additionally, the license includes the setup of initial prompts tailored to your specific use case as well as documentation and training to help your team effectively leverage the AI features.
🛠️ | ⚙️ | 🧱 | 💬 | 📚 |
Installation & Activation | General Configuration | Foundational Model Set-Up | Configuration of Prompts | Documentation & Training |
Considerations & Limitations
While our package offers robust capabilities, there are some limitations to be aware of. Professional service hours may be purchased to enhance the package to meet your specific requirements.
No Mogli Inbox support: Summary is configured to trigger from a record page or via a record-triggered flow. Please contact Mogli teams for customization options available through professional services.
Not all languages are supported: Generative AI models are trained using a dataset chosen by their source company and though some translation ability exists, performance of the summary will be impacted when used on languages where it was less well trained on.
Content of attachments are not summarized: Only the message text are used in the Summary, and contents within attachments such as PDF, video, audioclips, graphics and others will not be included in the summary.
Speciality knowledge may be limited: Generative AI models are trained using a dataset chosen by the Gen AI Service Provider. Your organization and clients may use brand names, trademarks, acronyms or specific terms-of-art that impact the quality of the summary.
Filters by related record: Conversation Summary and Message Recommendation will only use messages sent to, or received in response to messages sent from, the current record. While A contact and a case may share the same phone number, Conversation Summary will only summarize messages sent to, or received in response to messages sent from, the current record.
Websearch is excluded from the Summary process: Foundational models are trained on a certain dataset and will not be able to perform web searches in the course of Summary.
Additional choice of models and providers require separate scoping: The range of models available are dependent on our Gen AI service providers.
Summary of unsafe content / jailbreaking: Generative AI models are trained to reject requests to produce content which are considered dangerous. Conversation includes hate speech, sexually explicit material, self-harm may not be summarized. Attempts to circumvent safeties could result in the discontinuation of service.
Inclusion of information outside of messages: It is possible to further customize the Summary with other relevant information (eg. Address of the account, or status of a case). Please contact Mogli teams for additional customization options available through professional services.
Frequency of usage / throttling: Generative AI model providers have limits on number of requests that can be made within a period of time, requests beyond the allotment will have to be retried at a later date.
Notifications: It is possible for summaries to be emailed or added to chatter feed. Please contact Mogli teams for additional customization options available through professional services.
Alternative summary storage location: It is possible for summaries to be stored - either in a text field (text, long text area, rich text) on the parent record (so summaries can be overwritten) or in a text field of a child record of the parent record (such as task, so there can be multiple summaries on the opportunity). Please contact Mogli teams for additional customization options available through professional services.
While our package offers robust capabilities, there are some limitations to be aware of. Professional service hours may be purchased to enhance the package to meet your specific requirements.
Mogli URL and Scheduler links: Please contact Mogli teams for use cases that incorporate mogli URL and Scheduler links in the recommendation options available through professional services.
No bulk or batch processing: AI Message Recommendation is designed to be triggered one at time via the Mogli Conversation View.
Not all languages are supported: Generative AI models are trained using a dataset chosen by their source company and though some translation ability exists, performance of the suggestion will be impacted when used on languages where it was less well trained on.
Filters by related record: Message Recommendation will only use messages sent to, or received in response to messages sent from, the current record. While a Contact and a Case may share the same phone number, Message Recommendation will only be based on messages sent to, or received in response to messages sent from, the current record.
Content of attachments are not analysed: Only the message text are used in the recommendation, and contents within attachments such as PDF, video, audioclips, graphics and others will not be included in the summary.
Speciality knowledge may be limited: Generative AI models are trained using a dataset chosen by the Gen AI Service Provider. Your organization and clients may use brand names, trademarks, acronyms or specific terms-of-art that impact the quality of the summary.
Websearch is excluded from the recommendation process: Foundational models are trained on a certain dataset and will not be able to perform web searches in the course of recommendation.
Additional choice of models and providers require separate scoping: The range of models available are dependent on our GenAI service partners.
Generation of unsafe content / jailbreaking: Generative AI models are trained to reject requests to produce content which are considered dangerous. Conversation includes hate speech, sexually explicit material, self-harm may not be summarized. Prompts to create harmful recommendations will likely fail. Attempts to circumvent safeties could result in the discontinuation of service.
Inclusion of information outside of messages: It is possible to further customize the recommendation with other relevant information (eg. Address of the account, or status of a case). Please contact Mogli teams for customization options available through professional services.
Logging and ranking of recommendations are not included: It is possible to further customize this feature to log all recommendations for analysis and allow users to give feedback on the recommendations. Please contact Mogli teams for customization options available through professional services.
Frequency of usage / throttling: Generative AI model providers have limits on number of requests that can be made within a period of time, requests beyond the allotment will be refused.
Mogli AI Trust
Model Provider Claude by Anthropic Gemini by Google | |
Model Type Large Language Model | |
Model Input | Model Output |
|---|---|
| Generated summary of a conversation according to the focus and format as specified by the client. The outputs can be reviewed by the user |
Data used to train the AI system: None
Data used to provide Mogli AI: Mogli Message records (selected by the customer)
Personal Data: Personal data may be processed only if user inputs include personal data.
Data Retention: Mogli does not store your input and output after processing is complete. Usage data (how many times a feature is used) is stored in a secure, user restricted database.
Partner Data Retention:
Anthropic: 0 day retention
Gemini: 55 day retention
Anthropic Data Storage | Anthropic Commercial Terms | Gemini Terms
Your data is not used to train the AI system.