Frequently asked questions.

pilot5.ai is the platform that stops you from asking the same question to three different AIs and hoping for the best. It does two things. The assistant evaluates your question and either routes it to The Expert — the single best AI model for your specific domain, selected by benchmark — or activates a full deliberation, which puts five independent minds to work in a structured deliberation before you decide. One question. The right depth. Every time.

Deliberative AI is a decision-support approach where multiple AI models independently analyze a question, then cross-review each other's reasoning before a structured synthesis is produced. Unlike a single-model response that gives you one perspective, deliberation surfaces disagreements, blind spots, and minority viewpoints — so you make a better-informed decision. pilot5.ai applies deliberation only when your question warrants it. The assistant decides — and it never sends a simple question through five minds when one expert would have answered better.

Any single AI gives you one perspective, from one training run, with one set of blind spots. When you ask the same question to three different AIs, you get three answers — and you still have to decide which one to trust. pilot5.ai solves this differently. The assistant reads your question and either selects the single best expert automatically (The Expert), or assembles five independent minds that analyze, challenge each other, and converge on a structured recommendation. You get depth without overhead. One question in — one structured answer out.

The assistant is your conversational companion inside pilot5.ai. Before any deliberation, the assistant reads your question, evaluates its complexity, and recommends the right service — The Expert for focused questions, a full deliberation for decisions where multiple perspectives matter. The assistant never upsells. If your question needs one expert, it routes to The Expert. It will never push you toward a more expensive deliberation than your question requires. The assistant also narrates during deliberation and interprets the synthesis after — walking you through the confidence score, the decision matrix, and the Minority Report.

A deliberation assembles five independent minds — The Architect, The Strategist, The Engineer, The Counsel, and The Contrarian — each analyzing your question from a distinct angle. They work across multiple rounds: independent analysis, cross-critique, and final synthesis. Every dissent is preserved. You observe. You decide. The synthesis is the structured output — including a GO / PIVOT / STOP recommendation, confidence score, decision matrix, and Minority Report.

The Expert is a smart-routed single-model answer. pilot5 benchmarks 20+ models and selects the one that scores highest for your specific question domain — legal, financial, technical, strategic. Fast, affordable (avg ~0.2 credits), and ideal for straightforward questions.

When you want to guide the analysis yourself, you can. Choose the analytical lens:
— The Architect for structure, numbers, and operational clarity
— The Strategist for big picture, trends, and long-term framing
— The Engineer for technical precision and mathematical rigor
— The Counsel for risk assessment, legal nuance, and ethics
— The Contrarian to challenge your assumptions

And if you have a preferred model, you can specify it from our full roster of 20+ benchmarked experts. Most users leave it on automatic. The option is there when it matters.

The A-Team is the core deliberation mode. Five independent minds analyze your question in parallel (R1), then cross-critique each other's work anonymously (R2), across 2–4 adaptive rounds total. The orchestrator picks the round sequence based on what it detects in real time — tensions, confidence gaps, missing data — and can inject Devil's Advocate, Assumption Surfacing, Focused Research, or Calibration as additional rounds when the signals warrant it. The output is a structured recommendation — GO / PIVOT / NO GO (or INSUFFICIENT BASIS when grounding is too thin) — with a confidence score, an action plan, testable predictions, and a Minority Report.

The Dream Team is the most thorough mode. Everything in The A-Team, plus you in the room: deliberation pauses after every round to invite your steering, guidance, or correction — deterministic, never threshold-gated. It runs 4–6 adaptive rounds (vs 2–4 for A-Team), widens the calibration threshold so harder consensus questions trigger more depth, and is guaranteed at least one adaptive round (Devil's Advocate, Calibration, or Assumption Surfacing) before synthesis. Devil's Advocate also fires earlier — at agreement ≥ 0.8 vs ≥ 0.9 for A-Team. Best for board-level, capital-level, hiring-level decisions.

The Minority Report is the preserved dissenting view. When one of the five minds reaches a different conclusion from the majority, that perspective is not discarded — it is surfaced explicitly in the synthesis. A unanimous recommendation is not necessarily more reliable than one with dissent. It may simply mean the deliberation didn't surface a tension that was already there. The Minority Report ensures you always see what the majority might have missed. Read it before you decide.

The Expert typically responds in 8–18 seconds (single smart-routed model, no critique chain). The A-Team completes in roughly 6–9 minutes depending on round mix (2–4 adaptive rounds, no human pauses). The Dream Team typically runs 12–18 minutes with your input between rounds (4–6 adaptive rounds, deterministic HITL pause after each).

Yes. After every deliberation, pilot5.ai shows you which models were selected for each role. Fully transparent once it completes. You never see model names before the deliberation starts — the assistant selects in real time based on your question. But the full composition of your panel is always available in your deliberation history.

pilot5 uses a credit-based system where 1 credit = $1.00 USD. You only pay for actual AI compute used:

The Expert — avg ~0.5 credits (0.4–0.7 range)
The A-Team — avg ~4.2 credits (3–5.5 range)
The Dream Team — avg ~5.4 credits (3.8–7 range)

Cost is estimated before you confirm, debited on completion, and any excess is always refunded. See our Pricing page for plans and credit packs.

Credits are valid for 12 months from the date of purchase.

There is no free trial — every deliberation uses real AI compute. pilot5.ai is a professional tool, not a freemium product. The question framing — the step where the assistant reads your question and helps you frame it before any deliberation — is always free and uses no credits. For new users, the minimum entry is a $20 Starter credit pack — approximately 100 Expert answers or 10 full A-Team deliberations.

Yes. pilot5.ai is designed with GDPR in mind and built with EU AI Act readiness. Your deliberation content is never used to train AI models. All data in transit is encrypted via TLS, and our database enforces row-level security so you can only access your own data. Read our full Privacy Policy for details.

pilot5.ai benchmarks 20+ models selected for their complementarity, not their popularity. The platform dynamically selects the best model for each role based on your question's domain, required capabilities, and budget tier. After every deliberation, you can see exactly which models were used — full transparency.

The Confidence Score (0–10) measures three factors: agreement strength across the five minds, the ratio of verified evidence to inferred claims, and reasoning depth. A score of 8.2/10 means strong convergence with grounded evidence. A score of 5.4/10 means the panel was split or evidence was limited — read the Minority Report carefully before acting.

Yes. Your deliberation history is stored securely and becomes part of your personal knowledge layer — the deliberation engine can draw on past insights to give you richer context over time. You can delete any deliberation at any time from your history. Data is retained for up to 2 years unless you request earlier deletion.

pilot5.ai draws on five parallel knowledge layers: your deliberation memory, your uploaded documents, 250+ curated institutional sources (OECD, EU institutions, national regulatory authorities), live web intelligence from three independent search systems, and 100+ domain-specialist adapters across 21 knowledge domains. All institutional sources are primary — no aggregator noise.

Yes. pilot5.ai is built as a native MCP (Model Context Protocol) server with 12 tools: deliberate, check status, get results, get action plan, list history, search memories, search documents, search trusted sources, discover, submit feedback, resume deliberation, and check credits. Connect from any MCP-compatible AI client — Claude Desktop, Claude Code, Cursor, or your own tooling.

Regional panels are preset configurations that prioritize region-specific AI models. For example, an EU panel may prioritize European-origin models for data sovereignty, while an APAC panel selects models benchmarked for Asian-market expertise. This ensures your deliberation draws on models with the strongest regional knowledge.

Yes. You can update your regional preference at any time in your account Settings. The change takes effect on your next deliberation.