Platform Guides
Goals & Evaluation
Define measurable goals for your AI agents, track competencies over time, and use AI-powered evaluation to ensure consistent quality.
Overview
Flametalk's goal and evaluation system helps you measure and improve agent performance through two key capabilities: goal tracking to measure success rates, and competency evaluation to assess conversation quality.
Goal Tracking
Define measurable objectives and track success rates across calls
Competency Scoring
AI-powered scoring against defined competency criteria
Defining Agent Goals
Goals are the measurable outcomes you want your agent to achieve. Define clear, specific goals that can be objectively evaluated from call transcripts.
Types of Goals
Conversion Goals
- • Appointment booked
- • Demo scheduled
- • Sale completed
- • Lead qualified
Information Goals
- • Contact info collected
- • Survey completed
- • Feedback gathered
- • Issue documented
Resolution Goals
- • Problem resolved
- • Question answered
- • Complaint handled
- • Request fulfilled
Engagement Goals
- • Follow-up agreed
- • Interest confirmed
- • Referral obtained
- • Opt-in collected
Setting Up Goals
- 1
Navigate to your agent and click Goals.
- 2
Click + Add Goal.
- 3
Enter a clear, specific goal name (e.g., "Appointment Booked").
- 4
Describe the success criteria that indicates this goal was achieved.
- 5
Set a target percentage (e.g., 30% goal achievement rate).
# Goal: Appointment Booked ## DescriptionThe caller agrees to schedule an appointment with our team. ## Success Criteria- Caller provides a date/time preference OR- Agent confirms a specific appointment time OR- Cal.com booking action was triggered ## Target30% of answered calls should result in an appointment ## NotesThis is our primary conversion goal for sales calls.Competency Evaluation
Competencies measure how well your agent performs during conversations. Unlike goals (which are outcomes), competencies assess the quality of the conversation itself.
Competency Criteria
Define what "good" looks like for your calls:
| Competency | What It Measures | Weight |
|---|---|---|
| Greeting Quality | Professional introduction, name used correctly | 10% |
| Information Accuracy | Correct facts, no hallucinations | 25% |
| Problem Resolution | Addresses caller's main concern effectively | 30% |
| Tone & Empathy | Friendly, patient, appropriate tone | 15% |
| Call Closure | Clear next steps, proper sign-off | 20% |
Custom Competencies
You can define custom competencies specific to your use case. For example, compliance checks for regulated industries, brand voice alignment, or specific product knowledge verification.
Enabling Auto-Evaluation
- 1
Go to your channel settings.
- 2
Enable Auto Evaluation.
- 3
Set the sample rate (e.g., evaluate 100% of calls or a subset).
- 4
Configure your competency criteria and weights.
Understanding Scores
Each evaluated call receives a score from 0-100:
90-100
Excellent
70-89
Good
< 70
Needs Improvement
Analytics Dashboard
Track goal achievement and competency scores from the Analytics dashboard:
1,234
Total Calls
32%
Goal Achievement
87
Avg Competency
↑ 12%
vs Last Week
Goal Performance by Period
Manual Testing
Before deploying to production, test your agent manually to verify behavior:
Test Call
Make a real voice call to your agent directly from the dashboard. Great for quick verification and experiencing the agent as a caller would.
Dashboard → Agent → Test → Test Call
Chat Preview
Text-based chat with your agent (without voice). Faster iteration when testing conversation logic.
Dashboard → Agent → Test → Chat Preview
Best Practices
Start with One Goal
Focus on your most important outcome first. Once you achieve consistent results, add additional goals.
Make Goals Specific
"Book appointment" is better than "Successful call." Specific goals are easier to measure and optimize for.
Review Low Scores
Set up alerts for calls scoring below 70. Review these transcripts to identify improvement opportunities.
Iterate on Competencies
Refine your competency criteria based on what matters most. Start simple and add criteria as you learn what drives success.
Set Realistic Targets
Start with achievable targets and increase them as you improve. A 20-30% goal achievement rate is often a good starting point.
Continuous Improvement Loop
Goals and evaluation enable a continuous improvement cycle:
Deploy
Track Goals
Evaluate
Improve
- 1
Deploy your agent to production with defined goals.
- 2
Track goal achievement rates over time.
- 3
Evaluate call quality and identify patterns in low scores.
- 4
Improve agent instructions based on findings.