#Panel of Experts — Multi-AI Answers

Why ask one AI when you can ask three? Panel of Experts mode sends your question to multiple AI models simultaneously and lets you compare their answers side by side. It's like having a room full of consultants, except they don't charge by the hour and they never argue with each other. (Well, they might disagree — but politely.)

#What It Does

Normally, you pick one AI model and send it your question. Panel of Experts flips that on its head: instead of choosing a single provider, you send your question to two or three AI models at once — say, GPT-4, Gemini, and Claude — and get back separate responses from each. Every model receives the same context (your inventory, location, documents, conversation history), so the only difference is how each AI thinks about your question.

This is remarkably useful for important decisions, fact-checking, or simply getting diverse perspectives. One model might be more creative, another more precise, and a third more cautious. Seeing all three responses together gives you a richer, more balanced view than any single model can provide.

Each response is clearly labeled with the model that generated it, complete with a provider icon so you can tell at a glance who said what. No anonymous hot takes — full attribution, every time.

#How to Use It

#iOS App

#Enabling Panel Mode

1. Open the Chat tab
2. Tap the model name displayed near the top of the chat to open the Model Selector
3. Look for the Panel of Experts toggle
4. Switch it On
5. Select which models you want on your panel (choose 2 or 3 from the available providers)
6. Close the selector

#Asking a Question

1. Type your question in the chat text field, just like normal
2. Tap Send
3. Instead of one response, you'll see multiple response cards appear — one from each model on your panel
4. Each card shows:
   • The provider icon and model name (so you know who's talking)
   • The full response text
   • The same formatting, links, and lists you'd get from a single-model response
5. Scroll through to compare perspectives

#Switching Back to Single Mode

1. Tap the model name at the top of chat
2. Toggle Panel of Experts to Off
3. Select your preferred single model
4. You're back to one response per question

#Via Chat

Even in panel mode, you can ask meta-questions:

• "Which model gave the best answer to my last question?" (This goes to your currently selected single model)
• "Summarize the differences between the three responses."
• "Ask all three: what spices go best with salmon?"

#Via Claude Desktop (MCP)

The query_expert MCP tool lets you route questions to specific models from Claude Desktop:

• Use the provider parameter to choose: "openai", "google", "anthropic", or "perplexity"
• Use the model parameter for a specific model ID (e.g., "gpt-4o", "gemini-2.0-flash", "sonar-pro")
• The tool automatically includes your full ambientChat context with each query

Example: "Ask Perplexity what the current market value is for a 2023 MacBook Pro M3."

This isn't quite the same as Panel mode (you're asking one model at a time), but it gives you on-demand access to any provider's expertise without leaving Claude Desktop.

#Tips & Tricks

• Use Panel mode for high-stakes questions. Making a big purchase decision? Getting health-related advice? Trying to decide whether that cheese is still safe to eat? Multiple perspectives are genuinely valuable when the answer matters.
• Compare reasoning styles, not just conclusions. Sometimes all three models reach the same answer but get there differently. The reasoning path can be as informative as the final answer — one might cite your specific inventory data while another brings general knowledge.
• Switch back to single mode for casual questions. Panel mode uses tokens from multiple providers, which adds up. Save it for when you actually want multiple perspectives, and use single mode for "What's expiring this week?" and other routine queries.

#Options

#Known Limitations

• Higher token usage. Each model on the panel consumes tokens independently. A three-model panel uses roughly three times the tokens of a single-model query. This draws from your subscription credits faster, so keep an eye on your usage indicator.
• Some models are slower than others. Responses arrive as each model finishes, not all at once. You might see Gemini's answer in two seconds and Claude's in five. This is normal — different models have different processing speeds. The fast one isn't necessarily the right one.
• Not all subscription tiers support panel mode. Panel of Experts is a power feature that may require a higher subscription tier. If you don't see the toggle, check your subscription level in Account settings.
• Context is identical, but interpretation varies. All models receive the same inventory, location, and history context. Differences in responses come from each model's training, reasoning approach, and personality — not from different data. This is a feature, not a bug.
• No automatic consensus or "best answer" selection. The app presents all responses equally. It's up to you to decide which answer (or combination of answers) is most helpful. The app trusts your judgment, which is honestly the right call.

#Version History