Conversation Intelligence FAQs

Our Conversation Intelligence Dashboard uses contextual AI to surface insights from conversations. From emotional tone to conversation growth and trending topics, gain a better understanding of the impact you're having on your contacts as you help them toward more positive outcomes--then use this information to shape future outreach and communication strategy.

With that, we know there is a lot of information and questions that you're curious about.

This article will contain:

FAQs
Commonly used terms/definitions
A few extra pieces of information we think you'll want to know!

Frequently Asked Questions

Q. What defines a “conversation”?

A. A set of messages containing at least one incoming and one outgoing message prior to 3 days of inactivity.

Automated messages such as smart message auto-replies, keyword auto-replies, and OOO messages are analyzed as part of conversation data.

Q. Are smart messages included? What about keyword opt-outs?

A. Yes - both smart messages and keyword opt outs technically fit our definition of a “conversation” so they are part of the analysis.

Q. Who sees the Organization data v. Team v. Mine?

A. Admins see all, Users only see Mine.

Q. I can't click into the dashboard, how can I see my conversations?

A. Stay tuned! We'll be including additional enhancements in a future release this year. For now, your client success team can help provide that information. Use the "Schedule a Call" button on the dashboard to get yours scheduled!

Q. How can I drill down further/see specific conversations?

Q. Will WhatsApp be included in conversation data on the dashboard?

A. Stay tuned! It's not currently being calculated into the data but we do have plans to add WhatsApp information to the dashboard.

Q. Will Chat be included in conversation data on the dashboard?

A. Stay tuned! It's not currently being calculated into the data but we do have plans to add Chat information to the dashboard.

Q. How is my data being used?

A. We don’t train AI on your data or store it in a place where it can be accessed by AI. We simply use AI to help analyze data using very specific guidelines.

Q. What if multiple topics were discussed?

A. Our analysis only allows for one topic per conversation. If multiple are detected, the most relevant topic within the conversation will be selected.

Q. How do you actually know if a conversation was resolved or not?

A. Using AI, we perform a contextual analysis to identify the outcome of a conversation based on the messages within it.

Q. How do you calculate the sentiment shown?

A. We have a formula that takes into account Positive, Neutral, and Negative sentiment then arrives at an overall score that is visually expressed as a “bar” on the dashboard.

The analysis is instructed to only extract one emotion per conversation and it's based on only the contact's messages--not the user. If there are multiple, it'l select the most relevant/prevalent one.

Terms and Definitions

Conversation- One conversation = a set of messages containing at least one incoming and one outgoing message. After 3 days of inactivity, the conversation is considered "closed" and is analyzed.

Conversation Topic- A set of attributes extracted from message data in order to understand the main topics/subjects discussed by staff and their constituents.

Conversation Mood- A set of attributes extracted from message data in order to understand the emotional tone of staff and their constituents.

Joy

Anger

Sadness

Confusion

Satisfaction

Fear

Surprise

Interest

Neutral

Conversation Resolution- After a conversation has closed due to a 3-day period of inactivity, the AI will attempt to classify it as one of the following:

Resolved
Responded
Unresolved

If it is classified as "Unresolved" but the last message was an outgoing one, we classify the status as "Responded."

Conversation Sentiment- A determination of positive or negative tone within a set of messages from a contact. Ranges from Very Negative, Negative, Neutral, Positive and Very Positive.

Contact to Conversation Rate-The number of conversations (see definition above) that the contact has engaged in per contact.

Additional Information

Security and Privacy

The entirety of the technical implementation of this dashboard follows all existing security and privacy policies and procedures without exception. To reiterate that approach and touch on the specific use of AI models, those details are as follows:

All related application code and infrastructure is deployed entirely within Microsoft Azure in our existing SOC 2 Type II audited environment
All AI functionality is performed on private, internal-only, OpenAI GPT models hosted in this same environment
At no point is any institution data utilized for training or fine-tuning any AI models
At no point is any external, public AI model utilized for any functionality (e.g. ChatGPT, Grok, DeepSeek, etc)
At no point is any institution data stored in a location that is accessible to any public AI models.

All data processing, storage and ownership is consistent with our existing privacy and security policies and procedures, including

FERPA-compliance
Encryption of all data in transit and at rest
Institution retention of data ownership
Data deletion upon request and at contract termination

Feedback

We'll be continuing adding enhancements and data to this dashboard with future releases so stay tuned as to what is coming! But in the meantime, we'd love to hear what would be important for you to have on the dashboard!

Use the "Schedule a Call" option on the dashboard to schedule a review of your data with your client success team so we can help track what would help you and your teams best.