AI Setup and Configuration
Site Service Pro includes powerful AI capabilities that can analyze your project data, generate reports, answer questions, and provide intelligent insights. This guide covers everything you need to configure and optimize your AI features.
Understanding AI Modes
Site Service Pro offers two primary AI modes to meet different organizational needs:
Commercial AI (Cloud-Based)
Commercial AI uses industry-leading cloud AI providers to process your requests:
| Provider | Best For |
|---|---|
| OpenAI | General purpose, fast responses, excellent tool use |
| Anthropic | Complex analysis, detailed reports, nuanced reasoning |
| Google AI | Large context windows, document analysis |
| Azure OpenAI | Enterprise compliance requirements, existing Azure infrastructure |
| OpenAI-Compatible | Third-party AI services (OpenRouter, Together AI, Fireworks AI, etc.) |
| Custom / Self-Hosted | Full control with Ollama, LM Studio, LocalAI, and other local AI servers |
Advantages:
- Access to the latest AI models
- Regular model updates delivered automatically
- Web search capability (via Brave Search)
- No infrastructure management required
Considerations:
- Data is processed by the selected cloud provider
- Per-request API costs (billed directly to your provider account - check your provider's pricing)
- Requires internet connectivity
Private AI
Private AI runs through a centralized, managed AI gateway provided by Site Service Pro:
How it works:
- Your AI requests are processed through a dedicated, managed gateway
- Configuration is automatically synced from your Site Service Pro account
- No manual server configuration required — everything is managed for you
- Load balancing and failover are handled automatically
Advantages:
- Enhanced data handling through managed infrastructure
- No need to manage API keys for third-party providers
- Predictable experience with pre-configured models
- Automatic configuration updates
Considerations:
- Model selection is managed by Site Service Pro
- Web search is disabled in Private AI mode for data isolation
- Requires an active Private AI subscription
- Capabilities may differ from commercial models
Note: Private AI must be enabled for your account. Contact your administrator or Site Service Pro support to enable Private AI access.
Custom / Self-Hosted AI
For organizations that want to run AI models on their own infrastructure, Site Service Pro supports several self-hosted options:
| Provider Type | Description |
|---|---|
| Ollama | Popular local LLM runner, easy to set up |
| LM Studio | Desktop application for running local models |
| LocalAI | Privacy-focused local AI inference |
| Text Generation WebUI | Web-based interface for local models |
| KoboldAI | Specialized local inference engine |
| OpenAI-Compatible | Any server using the OpenAI API format |
Advantages:
- Complete control over your AI infrastructure
- Data never leaves your network
- Works in air-gapped environments
- No per-request API costs
Considerations:
- Requires dedicated hardware (GPU recommended)
- You manage updates and model deployment
- Performance depends on your hardware
- Some features may be limited depending on model capabilities
Initial AI Setup
Step 1: Access AI Configuration
- Click your profile picture in the top-right corner
- Select Account Settings from the dropdown
- Navigate to the AI Configuration section
Or for administrators: Administration → AI Configuration
Step 2: Select Your AI Mode
To use Commercial AI:
- Set AI Mode to "Commercial"
- Select your preferred AI Provider
- Enter your API Key for the selected provider
- Choose a Model from the available options
- Click Save Configuration
- Click Test Connection to verify
To use Private AI:
- Set AI Mode to "Private"
- Verify your connection status shows as active
- Configuration is managed automatically — no additional setup needed
- Click Save Configuration
To use Custom / Self-Hosted AI:
- Set AI Mode to "Commercial"
- Select Custom/Self-Hosted as your provider type
- Choose your Provider Type (Ollama, LM Studio, LocalAI, etc.)
- Enter your server's Endpoint URL
- Enter an API Key if your server requires one
- Enter the Model Name to use
- Click Save Configuration
- Click Test Connection to verify
Important: After changing AI configuration, we recommend testing with a simple query in AI Chat to confirm everything is working correctly.
Commercial AI Provider Setup
OpenAI Setup
- Visit OpenAI Platform
- Create an account or sign in
- Navigate to API Keys
- Click Create new secret key
- Copy the key and paste it in Site Service Pro
- Select your preferred model from the dropdown
Anthropic Setup
- Visit Anthropic Console
- Create an account or sign in
- Navigate to API Keys
- Generate a new API key
- Copy the key and paste it in Site Service Pro
- Select your preferred Claude model
Google AI Setup
- Visit Google AI Studio
- Create or sign in with your Google account
- Navigate to Get API Key
- Create a new API key
- Copy the key and paste it in Site Service Pro
- Select your preferred Gemini model
Azure OpenAI Setup
Azure OpenAI requires additional configuration:
- Create an Azure OpenAI resource in Azure Portal
- Deploy your desired model
- Copy your Endpoint URL
- Copy your API Key from Azure Portal
- Enter your Deployment Name in Site Service Pro
Note: Azure OpenAI is ideal for organizations with existing Azure infrastructure and enterprise compliance requirements.
OpenAI-Compatible Setup
For third-party services that use the OpenAI API format:
- Enter the Endpoint URL provided by your service
- Enter your API Key
- Enter the Model Name to use
- Click Test Connection to verify compatibility
This option works with services like OpenRouter, Together AI, Fireworks AI, and any other OpenAI-compatible API endpoint.
Web Search Configuration (Brave Search)
Commercial AI can search the web for current information using Brave Search:
Setting Up Web Search
- Visit Brave Search API
- Create an account and subscribe to a plan (free tier available)
- Copy your API Key
- In Site Service Pro, navigate to Account Settings → AI Configuration
- Click Configure Web Search
- Enter your Brave Search API Key
- Enable the Web Search toggle
- Click Test Connection to verify
- Click Save
When Web Search is Used:
- Questions about current events or real-time information
- Requests for up-to-date pricing, regulations, or news
- Looking up businesses, contact information, or websites
- Research questions requiring recent data
Note: Web Search is only available in Commercial AI mode. Private AI mode does not have internet access for data isolation purposes.
AI Configuration Options
Model Parameters
Fine-tune AI behavior with these settings:
| Parameter | Range | Default | Description |
|---|---|---|---|
| Temperature | 0.0 - 2.0 | 0.7 | Controls creativity. Lower = more focused, Higher = more creative |
| Max Output Tokens | 256 - 128,000 | 2,048 | Maximum response length |
| Context Limit | 2,048 - 200,000 | 4,096 | Maximum conversation context size |
| Top P | 0.0 - 1.0 | 1.0 | Nucleus sampling. Usually best left at default |
| Frequency Penalty | -2.0 to 2.0 | 0.0 | Reduces repetitive phrases |
| Presence Penalty | -2.0 to 2.0 | 0.0 | Encourages topic variety |
Recommended Settings by Use Case:
| Use Case | Temperature | Max Tokens | Notes |
|---|---|---|---|
| Data queries and reports | 0.3 - 0.5 | 4,096 | Lower temperature for accuracy |
| General conversation | 0.7 | 2,048 | Balanced default |
| Creative writing | 0.8 - 1.0 | 4,096 | Higher temperature for variety |
Feature Toggles
Enable or disable specific AI capabilities:
| Feature | Default | Description |
|---|---|---|
| AI Features (Master Switch) | Off | Must be enabled to use any AI features |
| AI Chat | On | General AI chat interface for querying data |
| Data Tools | On | Allows AI to query your project data (read-only) |
| RAG System | On | Smart skill matching for better query understanding |
| RFI Assistance | On | AI help within the RFI module |
| Timesheet Assistance | On | AI help within the timesheet module |
| Project Assistance | On | AI help within the project module |
| Document Analysis | On | AI-powered document extraction and analysis |
Tip: You can enable or disable individual features without affecting others. For example, you can use AI Reports while keeping AI Chat disabled.
Agentic Mode (Advanced)
Agentic mode allows the AI to perform multi-step reasoning — automatically chaining multiple data queries together to answer complex questions.
| Setting | Default | Description |
|---|---|---|
| Enable Agentic Mode | Off | Allow AI to perform multi-step queries automatically |
| Max Reasoning Steps | 5 | Maximum number of data query rounds per request (1-20) |
| Sequential Thinking | Off | AI shows its step-by-step reasoning process |
Example of Agentic Mode in Action:
You ask: "Compare labor costs across all active Houston projects"
The AI will automatically:
- Find all active projects in Houston
- Query timesheet data for each project
- Synthesize the comparison and present results
Tip: Start with agentic mode disabled. Enable it once you are comfortable with basic AI queries and want more powerful multi-step analysis.
Chat History Settings
| Setting | Default | Description |
|---|---|---|
| Include Chat History | On | Include previous messages for context |
| Max History Messages | 10 | Number of previous messages to include (1-50) |
Usage Limits
Optionally set limits on AI usage:
| Setting | Default | Description |
|---|---|---|
| Daily Request Limit | Unlimited | Maximum AI requests per day (0 = unlimited) |
| Hourly Request Limit | Unlimited | Maximum AI requests per hour (0 = unlimited) |
AI Health Status
Site Service Pro continuously monitors your AI connection:
- A health indicator shows whether your AI service is responsive
- If the connection fails, you will see the reason and when it was last working
- Health status is visible in the AI Configuration section
If your AI shows as unhealthy:
- Check that your API key is still valid
- Verify your provider account has available credits or quota
- Try the Test Connection button to diagnose the issue
- For Private AI, check that your subscription is active
Dynamic Model Catalog
Site Service Pro keeps your available AI models up to date automatically:
- New models appear in your selection dropdown as they become available
- Model details like capabilities and context window sizes are displayed
- Recommended models are highlighted to help you choose
- No software update is needed when new models are released
Troubleshooting
Common Issues
| Issue | Solution |
|---|---|
| "AI features not available" | Ensure AI Features master switch is enabled in settings |
| "Invalid API key" | Re-enter your API key; check it has not expired |
| "Connection failed" | Verify internet connectivity; check provider status page |
| "Rate limit exceeded" | Wait and try again; consider upgrading your provider plan |
| "Model not available" | Select a different model from the dropdown |
| Slow AI responses | Try a faster model; reduce max tokens; check your connection |
| Private AI not connecting | Verify your subscription includes Private AI; contact support |
Getting Help
If you need assistance with AI configuration:
- Contact your organization's administrator
- Reach out to Site Service Pro support
- Check the status page of your AI provider