Fix My Agent - Future AGI Documentation

After running simulations, Future AGI’s Fix My Agent feature automatically analyzes your agent’s performance and provides actionable recommendations to improve quality, reduce failures, and enhance overall effectiveness. Instead of manually debugging issues, get intelligent suggestions with one click.

Overview

Fix My Agent is your AI-powered diagnostic tool that turns simulation data into actionable insights. After each simulation run, the platform:

Analyzes simulation performance metrics and call patterns
Identifies specific issues and failure modes
Prioritizes recommendations by impact and urgency
Suggests targeted fixes you can implement immediately
Generates optimized system prompts automatically (optional)

Think of it as having an AI expert reviewing your agent’s conversations and telling you exactly what needs to be fixed—no manual debugging required.

Fix My Agent provides instant diagnostics and suggestions. For teams needing advanced prompt refinement, the platform also offers optimization algorithms (described later in this guide) that can automatically generate and test multiple prompt variations.

Quick Start: Recommended Workflow

⚡ Run simulation → Click “Fix My Agent” → Get instant suggestions
✏️ Implement fixes manually → Update your system prompt based on recommendations
✅ Validate → Re-run simulation to confirm improvements
🔄 Iterate → Repeat until your agent meets quality goals

95% of teams get great results with just steps 1-3. Auto-optimization is available if you need to test many prompt variations or want production-grade automated refinement.

Using Fix My Agent

After running a simulation, you can access Fix My Agent directly from the execution results page to get instant diagnostics and recommendations.

Step 1: Navigate to Simulation Results

Once your simulation run completes, you’ll see the execution details page with performance metrics including:

Call Details: Total calls, connected calls, connection rate
System Metrics: CSAT scores, agent latency, WPM (Words Per Minute)
Evaluation Metrics: Custom evaluation results

Step 2: Open Fix My Agent Panel

Click the “Fix My Agent” button in the top-right corner of the execution page. This opens a side panel showing:

All Suggestions: Total number of issues identified
Priority Levels: High, Medium, or Low priority for each issue
Issue Categories: Specific problems identified (latency, response brevity, detection tuning)
Affected Calls: Number of calls impacted by each issue
Last Updated: Timestamp of the latest analysis

Fix My Agent automatically analyzes your simulation results and generates suggestions by identifying patterns, edge cases, and failure modes. No configuration required—just click and get actionable recommendations.

Understanding Suggestions

Each suggestion provides:

Issue Description: Clear explanation of the identified problem
Recommended Fix: Specific action to address the issue
Priority Level: Urgency of the fix (High/Medium/Low)
Affected Calls: Which calls exhibited this issue
View Issue Button: Deep-dive into specific call examples

Example Suggestions:

Aggressively Reduce Pipeline Latency - Reduce LLM time-to-first-token (TTFT) by switching to a faster model
Enforce Strict Response Brevity - Implement a hard token limit to enforce concise responses
Tune End-of-Speech Detection - Adjust VAD parameters for better conversation flow

Start with High Priority suggestions that affect the most calls. These typically have the greatest impact on overall agent performance.

Advanced: Auto-Generate Optimized Prompts

After reviewing Fix My Agent suggestions, you have two options:

Implement suggestions manually - Take the recommendations and update your prompts yourself (recommended for most users)
Auto-generate optimized prompts - Use advanced optimization algorithms to automatically create and test multiple prompt variations

For teams that want automated prompt refinement, the platform includes powerful optimization algorithms that can systematically improve your agent’s system prompt.

Step 3: Configure Auto-Optimization (Optional)

If you want to automatically generate optimized system prompts, click the “Optimize My Agent” button in the Fix My Agent panel to open the optimization configuration dialog.

Required Configuration:

1. Name Your Optimization Run

Enter a descriptive name (e.g., “opt1”, “latency-optimization-v2”)
This helps track multiple optimization experiments

2. Choose Optimizer Select from Future AGI’s advanced optimization algorithms:

Random Search

Best for: Quick baseline testing and initial explorationHow it works: Generates random prompt variations using a teacher model and evaluates each candidate.Characteristics:

⚡⚡⚡ Fast execution
⭐⭐ Basic quality improvements
💰 Low cost
Ideal for: 10-30 examples

Use when: You need quick results or want to establish a performance baseline before trying more sophisticated algorithms.

Bayesian Search

Best for: Few-shot learning tasks and intelligent example selectionHow it works: Uses Bayesian optimization to intelligently select few-shot examples and prompt configurations.Characteristics:

⚡⚡ Medium speed
⭐⭐⭐⭐ High quality
💰💰 Medium cost
Ideal for: 15-50 examples

Use when: Your dataset contains good examples and you want to leverage few-shot learning effectively.

Meta-Prompt

Best for: Complex reasoning tasks requiring deep analysisHow it works: Analyzes failed examples, formulates hypotheses, and rewrites the entire prompt through deep reasoning.Characteristics:

⚡⚡ Medium speed
⭐⭐⭐⭐ High quality
💰💰💰 Higher cost
Ideal for: 20-40 examples

Use when: Your agent handles complex reasoning tasks or you need holistic prompt redesign.

ProTeGi

Best for: Identifying and fixing specific error patternsHow it works: Generates critiques of failures and applies targeted improvements using beam search to maintain multiple candidates.Characteristics:

⚡ Slower execution
⭐⭐⭐⭐ High quality
💰💰💰 Higher cost
Ideal for: 20-50 examples

Use when: You have clear failure patterns and want systematic error fixing.

PromptWizard

Best for: Creative exploration and diverse prompt variationsHow it works: Combines mutation with different “thinking styles”, then critiques and refines top performers.Characteristics:

⚡ Slower execution
⭐⭐⭐⭐ High quality
💰💰💰 Higher cost
Ideal for: 15-40 examples

Use when: You want creative exploration or diverse conversational approaches.

GEPA (Genetic-Evolutionary Prompt Algorithm)

Best for: Production deployments requiring state-of-the-art performanceHow it works: Uses evolutionary algorithms with reflective learning and mutation strategies inspired by natural selection.Characteristics:

⚡ Slower execution
⭐⭐⭐⭐⭐ Excellent quality
💰💰💰💰 Highest cost
Ideal for: 30-100 examples

Use when: You need production-grade optimization with robust results and have sufficient evaluation budget.

3. Select Language Model Choose the model that will be used for the optimization process: Available models include:

gpt-5 series (gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-chat-latest)
gpt-4 series (gpt-4, gpt-4.1, gpt-4o, gpt-4o-audio-preview)
Other supported models from your configuration

For optimization, using a more powerful model (like gpt-4 or gpt-5) as the teacher model often yields better prompt improvements, even if your production agent uses a smaller model.

4. Add Parameters Configure optimizer-specific parameters:

Number Variations: How many prompt variations to generate and test
- Start with 3-5 for quick iterations
- Use 10-20 for thorough optimization
- Consider cost vs. quality tradeoff

Each optimizer may have additional parameters. The platform shows recommended defaults that balance speed and quality.

Step 4: Start Auto-Optimization

Click “Start Optimizing your agent” to begin the automated prompt generation process. The optimization engine will:

Analyze your simulation data and Fix My Agent suggestions
Generate multiple system prompt variations using the selected algorithm
Evaluate each variation against your test scenarios
Score performance improvements
Select the best-performing optimized prompt

Most users find that manually implementing Fix My Agent suggestions is the fastest path to improvement. Use auto-optimization when you need to test many prompt variations or want production-grade automated refinement.

Advanced: Auto-Optimization Algorithms

For teams that choose to use automated prompt generation, Future AGI provides advanced optimization algorithms. This section explains how each algorithm works to help you choose the right strategy.

Most teams get excellent results by implementing Fix My Agent suggestions manually. These algorithms are for advanced use cases where you need to test many prompt variations automatically.

Quick Selection Guide

Your Goal	Recommended Algorithm	Why
Quick improvement baseline	Random Search	Fast, simple, establishes performance floor
Reduce latency issues	Bayesian Search	Efficiently explores configuration space
Fix conversation logic errors	ProTeGi or Meta-Prompt	Targets specific failure patterns
Improve complex reasoning	Meta-Prompt	Deep analysis and systematic refinement
Optimize for production	GEPA	State-of-the-art evolutionary optimization
Explore creative approaches	PromptWizard	Diverse variations with structured refinement

Algorithm Comparison

Algorithm	Speed	Quality	Cost	Best Dataset Size
Random Search	⚡⚡⚡	⭐⭐	💰	10-30 examples
Bayesian Search	⚡⚡	⭐⭐⭐⭐	💰💰	15-50 examples
Meta-Prompt	⚡⚡	⭐⭐⭐⭐	💰💰💰	20-40 examples
ProTeGi	⚡	⭐⭐⭐⭐	💰💰💰	20-50 examples
PromptWizard	⚡	⭐⭐⭐⭐	💰💰💰	15-40 examples
GEPA	⚡	⭐⭐⭐⭐⭐	💰💰💰💰	30-100 examples

Speed: ⚡ = Slow, ⚡⚡ = Medium, ⚡⚡⚡ = Fast
Quality: ⭐ = Basic, ⭐⭐⭐⭐⭐ = Excellent
Cost: 💰 = Low, 💰💰💰💰 = High (based on API calls)

Decision Tree

Do you need production-grade optimization?
├─ Yes → Use GEPA
└─ No
   │
   Do you have clear error patterns to fix?
   ├─ Yes → Use ProTeGi
   └─ No
      │
      Is your task reasoning-heavy or complex?
      ├─ Yes → Use Meta-Prompt
      └─ No
         │
         Do you need few-shot learning optimization?
         ├─ Yes → Use Bayesian Search
         └─ No
            │
            Do you want creative exploration?
            ├─ Yes → Use PromptWizard
            └─ No → Use Random Search (baseline)

Viewing and Deploying Improvements

For Manual Implementations

After implementing Fix My Agent suggestions:

Re-run simulations with your updated prompt
Compare metrics to baseline in the execution dashboard
Review new suggestions from Fix My Agent
Iterate until performance meets your goals
Deploy to production when satisfied

For Auto-Optimization Results

If you used automated optimization, view results in the Optimization Runs tab:

Performance Comparison
- Original prompt baseline scores
- Auto-generated prompt scores
- Improvement percentage
Best Prompt
- The highest-performing variation
- Changes made from the original
- Evaluation scores across metrics
Optimization History
- All variations tested
- Performance trajectory
- Iteration details

Deployment Checklist

Whether implementing manually or using auto-optimization: ✓ Review the improved prompt carefully
✓ Test with additional scenarios not in original dataset
✓ Update your agent definition with the new prompt
✓ Re-run simulations to validate improvements
✓ Monitor performance in production

Always validate with new test cases before production deployment. Both manual and automated approaches can overfit to the evaluation dataset.

Best Practices

1. Start with Fix My Agent Suggestions

Always begin with manual implementation:

Review all Fix My Agent suggestions after each simulation
Implement high-priority fixes first (greatest impact)
Re-run simulation to validate improvements
Only use auto-optimization if you need to test many variations

Fix My Agent provides instant, actionable recommendations that you can implement in minutes. Most teams see significant improvements by simply following the suggestions without needing automated optimization.

2. Use Sufficient Test Data

Fix My Agent works best with comprehensive simulation data:

Run at least 20-50 simulation scenarios before analyzing
Ensure scenarios cover diverse situations and edge cases
Include examples of both successful and failed interactions
More data = more accurate diagnostics

3. Implement Iteratively

Don’t try to fix everything at once:

Address 1-2 high-priority issues per iteration
Re-run simulations after each change
Verify improvements before moving to next issue
Track what worked and what didn’t

4. Use Auto-Optimization Strategically

If you choose to use automated optimization algorithms:

Latency issues: Bayesian Search (efficient parameter tuning)
Conversation logic errors: ProTeGi (targeted error fixing)
Complex reasoning: Meta-Prompt (deep analysis)
Production deployment: GEPA (robust evolutionary search)

5. Balance Cost and Quality

For auto-optimization (API calls required):

Start with fewer variations (3-5) for quick iterations
Increase variations (10-20) when you’re close to deployment
Use faster algorithms (Random Search, Bayesian Search) for experimentation
Reserve expensive algorithms (GEPA, Meta-Prompt) for critical optimizations

6. Always Validate Improvements

Whether implementing manually or using auto-optimization:

Run new simulations after making changes
Compare metrics against the baseline
Test on scenarios not included in the original dataset
Monitor for unexpected behaviors or regressions

Complete Workflow Example

Here’s how to improve an insurance sales agent using Fix My Agent:

Initial State

Agent has 40% call connection rate
High latency (1470ms response time)
Mixed sentiment scores

Step 1: Run Comprehensive Simulations

- Create 50 diverse scenarios covering:
  ✓ Different customer types
  ✓ Various objection patterns  
  ✓ Edge cases and difficult situations
- Run simulation and analyze results

Step 2: Open Fix My Agent

Click "Fix My Agent" button to get instant diagnostics

Suggestions identified:
- [High Priority] Reduce Pipeline Latency (8 calls affected)
  → Switch to faster model or reduce system prompt verbosity
  
- [High Priority] Enforce Response Brevity (8 calls affected)  
  → Add explicit instruction: "Keep responses under 50 words"
  
- [Medium Priority] Tune End-of-Speech Detection (8 calls affected)
  → Adjust endpointing delay parameters

Step 3: Implement High-Priority Fixes

Manual changes made to system prompt:
✓ Added: "Be extremely concise. Maximum 2 sentences per response."
✓ Switched model: gpt-4o → gpt-4o-mini (faster)
✓ Removed: Verbose examples from system prompt

Step 4: Validate Improvements

- Run new simulation with updated prompt
- Compare results:
  Before: 40% connection rate, 1470ms latency
  After: 65% connection rate, 850ms latency
  Improvement: +62.5% connection rate, -42% latency

Optional Step 5: Auto-Optimization (If Needed)

If manual fixes aren't sufficient, use auto-optimization:
- Name: "insurance-agent-production-v1"
- Optimizer: GEPA
- Model: gpt-4o
- Variations: 15
- Result: Additional 5% improvement in conversion rate

In this example, Fix My Agent provided instant, actionable suggestions that the team implemented in 10 minutes, resulting in 62.5% improvement. Auto-optimization was used as a final refinement step for production deployment.

Troubleshooting

No Suggestions in Fix My Agent

Possible causes:

Not enough simulation data (need 20+ calls)
Agent performed perfectly (no issues detected)
Evaluation metrics not configured

Solutions:

Run more comprehensive simulations
Add diverse scenarios including edge cases
Configure custom evaluation metrics to measure quality

Manual Fixes Not Improving Performance

Possible causes:

Suggestions not fully implemented
Changes introduced new issues
Need more comprehensive refinement

Solutions:

Double-check all high-priority suggestions are addressed
Test changes incrementally (one at a time)
Consider using auto-optimization for systematic refinement

Auto-Optimization Not Improving Performance

Possible causes:

Insufficient training data
Wrong optimizer for the problem type
Too few variations tested
Overfitting to evaluation set

Solutions:

Ensure you have 30+ diverse simulation scenarios
Try a different optimization algorithm (see selection guide)
Increase number of variations (10-20)
Validate on held-out test scenarios

Auto-Optimization Taking Too Long

Possible causes:

Using slow optimizer (GEPA, ProTeGi)
Too many variations configured
Large dataset size

Solutions:

Consider implementing Fix My Agent suggestions manually instead
Start with Random Search or Bayesian Search for faster results
Reduce number of variations to 3-5
Use a smaller sample of representative scenarios

Advanced Topics

Combining Fix My Agent with Auto-Optimization

Get the best of both worlds:

Use Fix My Agent to get instant diagnostic suggestions
Implement high-priority fixes manually for quick wins
Run auto-optimization for additional systematic refinement
Compare results between manual and automated approaches
Deploy the best-performing version

This hybrid approach is ideal for production deployments: get 80% improvement from manual fixes in minutes, then use auto-optimization to squeeze out the remaining 20%.

Custom Evaluation Metrics

Fix My Agent and optimization work better with custom evaluation metrics that match your business goals:

Conversion Rate: Did the agent successfully convert the customer?
Compliance: Did the agent follow regulatory requirements?
Customer Satisfaction: Sentiment and CSAT scores
Efficiency: Response latency, call duration, token usage

Both Fix My Agent diagnostics and optimization algorithms use your evaluation metrics to identify issues and measure improvements. Better metrics lead to better suggestions.

Fix My Agent for Different Agent Types

Different agent types see different patterns in their suggestions: Voice Agents:

Common issues: Latency, verbosity, interruption handling
Typical suggestions: Switch to faster models, reduce response length, adjust endpointing
Auto-optimization: Bayesian Search (parameter tuning), ProTeGi (error fixing)

Chat Agents:

Common issues: Response quality, accuracy, context retention
Typical suggestions: Improve instruction clarity, add examples, enhance context handling
Auto-optimization: Meta-Prompt (reasoning), PromptWizard (diverse styles)

Sales Agents:

Common issues: Conversion rate, objection handling, compliance
Typical suggestions: Better objection responses, clearer value props, compliance checks
Auto-optimization: GEPA (production-grade), Meta-Prompt (complex logic)

Support Agents:

Common issues: Problem resolution, response time, escalation logic
Typical suggestions: Clearer troubleshooting steps, empathy improvements, faster responses
Auto-optimization: ProTeGi (error patterns), Bayesian Search (few-shot examples)

Next Steps

Run Simulation

Learn how to run comprehensive agent simulations

Create Scenarios

Build diverse test scenarios for better diagnostics

Agent Definition

Configure your agent for optimal performance

Optimization Algorithms (Advanced)

Deep dive into auto-optimization algorithm details

Getting Started:

Run Your First Simulation - Start getting Fix My Agent suggestions
Create Test Scenarios - Build comprehensive test coverage
Evaluation Metrics - Configure better diagnostics

Advanced Auto-Optimization:

Prompt Optimization Overview - Learn about the agent-opt library
GEPA Algorithm - Evolutionary optimization deep dive
Meta-Prompt Algorithm - Deep reasoning refinement
ProTeGi Algorithm - Error-driven improvement

Get Started

Guides

​Overview

​Quick Start: Recommended Workflow

​Using Fix My Agent

​Step 1: Navigate to Simulation Results

​Step 2: Open Fix My Agent Panel

​Understanding Suggestions

​Advanced: Auto-Generate Optimized Prompts

​Step 3: Configure Auto-Optimization (Optional)

​Required Configuration:

​Step 4: Start Auto-Optimization

​Advanced: Auto-Optimization Algorithms

​Quick Selection Guide

​Algorithm Comparison

​Decision Tree

​Viewing and Deploying Improvements

​For Manual Implementations

​For Auto-Optimization Results

​Deployment Checklist

​Best Practices

​1. Start with Fix My Agent Suggestions

​2. Use Sufficient Test Data

​3. Implement Iteratively

​4. Use Auto-Optimization Strategically

​5. Balance Cost and Quality

​6. Always Validate Improvements

​Complete Workflow Example

​Initial State

​Step 1: Run Comprehensive Simulations

​Step 2: Open Fix My Agent

​Step 3: Implement High-Priority Fixes

​Step 4: Validate Improvements

​Optional Step 5: Auto-Optimization (If Needed)

​Troubleshooting

​No Suggestions in Fix My Agent

​Manual Fixes Not Improving Performance

​Auto-Optimization Not Improving Performance

​Auto-Optimization Taking Too Long

​Advanced Topics

​Combining Fix My Agent with Auto-Optimization

​Custom Evaluation Metrics

​Fix My Agent for Different Agent Types

​Next Steps

Run Simulation

Create Scenarios

Agent Definition

Optimization Algorithms (Advanced)

​Related Resources

Overview

Quick Start: Recommended Workflow

Using Fix My Agent

Step 1: Navigate to Simulation Results

Step 2: Open Fix My Agent Panel

Understanding Suggestions

Advanced: Auto-Generate Optimized Prompts

Step 3: Configure Auto-Optimization (Optional)

Required Configuration:

Step 4: Start Auto-Optimization

Advanced: Auto-Optimization Algorithms

Quick Selection Guide

Algorithm Comparison

Decision Tree

Viewing and Deploying Improvements

For Manual Implementations

For Auto-Optimization Results

Deployment Checklist

Best Practices

1. Start with Fix My Agent Suggestions

2. Use Sufficient Test Data

3. Implement Iteratively

4. Use Auto-Optimization Strategically

5. Balance Cost and Quality

6. Always Validate Improvements

Complete Workflow Example

Initial State

Step 1: Run Comprehensive Simulations

Step 2: Open Fix My Agent

Step 3: Implement High-Priority Fixes

Step 4: Validate Improvements

Optional Step 5: Auto-Optimization (If Needed)

Troubleshooting

No Suggestions in Fix My Agent

Manual Fixes Not Improving Performance

Auto-Optimization Not Improving Performance

Auto-Optimization Taking Too Long

Advanced Topics

Combining Fix My Agent with Auto-Optimization

Custom Evaluation Metrics

Fix My Agent for Different Agent Types

Next Steps

Related Resources