Aligning AI Insights to Your Business Goals
Voxjar doesn't generate generic conversation topics; your AI learns from your calls and can be tuned to understand what matters specifically to your business.
This guide will show you how to teach your AI about your company, your goals, and your definition of a "good call", transforming standard analysis into personalized, actionable intelligence.
How AI Insights Work
Every time you run an AI Evaluation or Transcribe an interaction, Voxjar automatically generates a rich set of AI Insights. This is included with every AI analysis and does not consume extra credits.
Your AI analyzes each interaction to produce a detailed breakdown, which you can find throughout the Voxjar application in tables, reporting and on any interaction's details page.
Image description: AI Insights by Voxjar, displaying the different analysis modules generated for a single call.
The Key: Teach your AI About Your Business
The single most important step to getting high-quality, relevant insights is to tell your AI what to look for.
You do this by in three ways:
A Company Prompt
Custom Topics
Custom Skills
1. Your Company Prompt
This is the high-level guidance for your AI. By providing a detailed description of your company, products, customers, and core values, you give the AI the foundational knowledge it needs to understand the "why" behind your interactions. This prompt acts as the strategic lens through which your AI views every conversation.
How to Write an Effective Company Prompt
Your Company Prompt is the most important piece of instruction you can give your AI. A well-written prompt directly leads to more relevant summaries and more accurate coaching suggestions. Use the following guide to craft a prompt that will give you the best results.
1. Your Company & Product/Service Description
Start with the basics. Who are you and what do you sell or do?
Example: "We sell accounting software designed specifically for freelance photographers and designers. Our main features are invoicing, expense tracking, and tax preparation."
2. Your Customer & Brand Voice
Who are your customers? What is your relationship with them like? What tone should your agents use?
Example: "Our customers are typically small business owners who are not accounting experts. Our brand voice is friendly, patient, and supportive. We avoid overly technical accounting terms."
3. The Primary Goal(s) of Your Interactions
This is critical. Tell your AI what you are trying to achieve.
Example: "The primary goal of a support call is First-Call Resolution. We want to solve the customer's problem completely on the first try so they don't have to call back."
4. Define "Success": Key Priorities & Behaviors
Get specific about what an ideal interaction looks like. List the positive behaviors you want your AI to identify as strengths.
Example: "We prioritize the following behaviors:
Empathy: Acknowledge the customer's frustration when they report a problem.
Clarity: Confirm the customer's issue by repeating it back to them before offering a solution."
5. (Optional) Define "Failure": What to Avoid
You can also tell your AI what you don't want to see.
Example: "Agents should avoid promising new features by a specific date."
How to set it:
Navigate to Settings > Branding.
Fill out the Company Description and Goals text box using the guide above.
Click Save Changes.
2. Categorize the "What" with Custom Topics
Topics are the subjects or reasons for an interaction. They identify what the conversation was about. The AI analyzes the transcript to identify these subjects and timestamps each topic. This transforms a simple label into a powerful navigational tool, allowing you to click and jump directly to the relevant moments in the conversation.
Primary Purpose:
Categorize Interactions: Group all calls about a specific subject, like "Billing Disputes" or "Product Feature Inquiries."
Identify Trends: Discover if certain subjects are being discussed more or less frequently over time. For example, you might see a spike in the "Shipping Damage" topic, indicating a problem with your carrier.
Navigate Calls Faster: Instead of just knowing a call was about a billing dispute, you can instantly listen to the exact part of the call where the billing issue was discussed.
How to manage them: Go to Settings > Topics to add your own topics, manage the default list, and review AI suggestions.
AI Suggestions:
Over time, your AI will identify relevant topics that are not currently being tracked. These suggestions can be added via your Topics settings page or you can enable AI recommendation emails via your Notifications Settings.
3. Evaluate the "How" with Custom Skills
Skills are the behaviors, competencies, and techniques demonstrated by your agents. They evaluate how the agent handled the interaction. Your AI uses your defined list of skills to identify and populate the Strengths and Weaknesses sections for each interaction. The AI doesn't just identify a skill; it provides a written explanation of how that skill was (or was not) demonstrated and attempts to timestamp the specific moment, giving you a concrete, data-backed example for coaching.
Primary Purpose:
Evaluate Performance: Assess agent effectiveness based on a standardized set of behaviors, such as "Active Listening" or "Empathy."
Provide Targeted Coaching: Pinpoint specific, actionable areas for improvement. The AI uses your Skills list to generate the "Strengths" and "Weaknesses" in each AI Insight report.
Identify Best Practices: Find call recordings where key skills were demonstrated exceptionally well and use them as training examples for the rest of your team.
How to manage them: Go to Settings > AI Settings > Skills to add your own skills, manage the default list, and review AI suggestions.
Learn More & Take Control
AI Suggestions:
Over time, your AI will identify relevant skills that are not currently being tracked. These suggestions can be added via your Skills settings page or you can enable AI recommendation emails via your Notifications Settings.
What's Inside an AI Insight?
Once you've set your company prompt, here's the kind of tailored analysis you can expect on every call:
Structured Summary: More than a simple paragraph, this is a clear breakdown of the call's Reason, Key Issues, Customer Pain Points, and the final Resolution.
Outcomes: The AI identifies the most critical results of the conversation. (e.g., "Potential $1,000,000 order secured," "Customer dissatisfaction expressed").
Strengths & Weaknesses: The AI assesses the agent's performance against your defined Skills (like Salesmanship, Empathy, Problem Solving) and highlights specific moments in the call as examples.
Missed Opportunities: Did the agent miss a chance to upsell, build rapport, or resolve an issue more efficiently? The AI flags these for targeted coaching.
Topics: The AI tags the conversation with key subjects discussed, complete with timestamps so you can jump right to the relevant part of the call.
Sentiment and Call Dynamics: Get a top-level view of the call's mood, talk-to-listen ratios, crosstalk, and periods of silence.
Use Cases: Putting AI Insights to Work
AI Insights are designed for speed, discovery, and autonomous analysis. They excel at giving you a high-level understanding of your interactions without requiring deep configuration. Here are a few ways to leverage the unique strengths of AI Insights.
Use Case 1: Instantly Triage and Find Your Most Critical Calls
You have hundreds of calls, but you only have time to review a few. How do you find the ones that need your attention the most? AI Insights provides objective data points to filter and find these "needle in a haystack" interactions.
How: Go to the Interactions page and create a filter that combines the metrics automatically generated by Insights. For example:
Customer Sentiment is Negative
AND
Crosstalk Duration is Greater than 15 seconds
Result: You have instantly isolated calls that were not just negative, but also contentious and difficult. Instead of listening to calls at random, you are focused on the interactions that offer the most significant coaching or customer-rescue opportunities.
Use Case 2: Discover Emerging Business Trends with Topics
Is a new product issue bubbling up? Are customers suddenly asking about a competitor? The Topics feature is a powerful way to get a bird's-eye view of what's happening on your front lines.
How: Use the Reporting dashboard to view your top Topics over the last 30 days. Watch for new topics appearing or existing ones climbing the ranks.
Result: You notice the topic "Invalid Coupon Code" has jumped from #10 to #2 this week. This isn't a support agent issue; it's a signal that your marketing department may have sent out a broken code. You've just used high-level conversation data to solve a core business problem, no deep analysis required.