We’re excited to announce the release of AI Questions: a new generative AI feature in Creovai Conversation Intelligence.
AI Questions use GenAI prompts to interrogate large sets of contact center interactions, acting as your conversation detective. It provides instant answers to questions that would have previously required you to comb through transcripts or listen to full call recordings, significantly reducing evaluation time and delivering information to help you make smart decisions and improve call outcomes.
How can you use AI Questions?
AI Questions analyze interactions, identify patterns, and generate answers based on the full context of a conversation, allowing you to uncover insights that go beyond traditional keyword-based analysis.
AI Questions are customizable, and there are many potential uses, including:
- Capturing call reasons—Manual call dispositioning is prone to human error and doesn’t always capture the full nuance of the call reason (for example, a customer might bring up multiple issues, and the agent is only able to select one disposition option). AI Questions help you understand the real reasons customers are calling, allowing you to improve self-service, make informed staffing decisions, and solve more issues upstream.
- Detecting anomalies—AI Questions can help you uncover anomalies that you might have missed otherwise, such as unexpected customer reactions or edge cases that are becoming more frequent. This enables you to address emerging issues before they become widespread problems.
- Predicting future customer needs or churn risk—You can run AI Questions to help determine if a customer’s needs were fully met, if the contact center needs to take any follow-up actions, if a customer is at risk of churning, and more. This can help you proactively resolve customer issues, reduce repeat contacts, and save customers before they churn.
- Determining the impact of agent behaviors—AI Questions can help you identify instances of agent behaviors that would be difficult to track through keyword-based analysis alone. You can also use AI Questions to determine how different agent behaviors are impacting outcomes, such as customer satisfaction and first-call resolution. This can help you improve agent coaching and real-time guidance, increasing the likelihood of positive call outcomes.
How do AI Questions work?
AI Questions use generative AI prompts that run at the interaction level in Creovai Conversation Intelligence. You can choose to run a prompt on all the interactions stored in the platform or on a large set of calls based on your criteria. You can tailor sets of AI Questions to specific interactions by applying filters, and the questions will run on all interactions that match your filters as the interaction data enters the Creovai platform.
You can choose between two different question types for your prompts:
- Open-ended—This will return an open-text response
- Multiple choice—This lets you provide a list of possible answers, with AI Questions applying one answer per interaction
You can benefit from AI Questions even if you don’t have a lot of experience with GenAI prompt engineering. We have a growing library of default prompts to help you get started, and our Customer Success team can also work with you to develop custom prompts that meet your needs.
What kind of questions can you answer?
We designed AI Questions to help you answer nuanced questions that require a full understanding of everything the agent and customer said in context. While our machine learning insight categories are great for answering objective questions about specific language usage (e.g., “Did the agent read the disclosure statement?”), AI Questions can generate responses based on the whole interaction.
Examples include:
- What was the reason for the call?
- Was the agent friendly (based on provided examples of friendly language)?
- What steps did the agent take to troubleshoot the customer's issue?
- What was the outcome of the interaction?
- Was the customer satisfied or dissatisfied by the end of the interaction?
- Why did the customer say they wanted to cancel?
- Why did the customer decide not to cancel?
How do AI Questions work alongside machine learning models?
Creovai takes a hybrid approach to AI, combining generative AI capabilities with machine learning models to 1.) understand conversation context and nuance and 2.) track objective utterance-based criteria, validate conversation insights, and quantify the impact.
You can use AI Questions alongside our proprietary machine learning models for a more complete understanding of your contact center interactions. Generative AI allows you to answer interaction-level questions in which context is critical. Machine learning lets you track objective utterance-based criteria (e.g., agents giving disclosure statements) across 100% of your interactions. Our machine learning models can also help you validate patterns in your interactions using confidence intervals, correlation analysis, and predictive modeling. Together, our AI models strengthen your understanding of agent performance, customer experience, and conversation trends so you can make data-backed business decisions.
How AI Questions fuel smarter decision-making
Manual call evaluations can be incredibly time-consuming, especially when evaluators have to listen to a lengthy call (sometimes multiple times) to find answers to their questions. While QA automation helps reduce call review time by automatically scoring objective criteria, answering nuanced questions at scale has always been a challenge—until now.
AI Questions help you identify the important events and behaviors happening in your contact center conversations, saving you time on manual evaluations and giving you valuable insights to inform your decision-making. You can:
- Reduce errors in documentation—Manual call summarization, dispositioning, and evaluation are all subject to human error and biases. AI Questions give you responses based on interaction transcripts, increasing your documentation accuracy and giving you better visibility into call reasons, agent behaviors, customer sentiment, and more.
- Understand any interaction at a glance—Supervisors don't have the bandwidth to sit on every call (or review every interaction), but AI Questions allow them to jump into any interaction record and quickly understand what happened. This can help improve coaching sessions and inform follow-up actions (such as outreach to dissatisfied customers).
- Answer questions to inform your contact center strategy—Combining AI Questions with the statistical validation of our machine learning models helps you understand what’s happening in your interactions and uncover your top opportunities for improvement. This can help you plan, prioritize, and quantify the impact of initiatives to reduce operational costs, decrease churn, increase sales, and improve the customer experience.
Our new AI Questions feature is now available in the Creovai Conversation Intelligence platform. If you’d like to learn more, you can request a demo now.
