Autopilot
Autopilot responds to every incoming fan message in your persona. This page covers strategy selection, rules, escalation, approval modes, and every advanced control.
Fanvue automated messages and autopilot
Do I still need to set up automated messages in Fanvue?
Yes. Fanvue's native automated messages (welcome messages to new subscribers, scheduled broadcasts) are completely separate from FanWake and run independently. FanWake autopilot only activates when a fan sends you a message — it does not send first messages to new subscribers on its own.
This means your Fanvue welcome message is critical: it sparks the first reply that autopilot then handles. A weak welcome message that fans ignore means autopilot never gets triggered for that subscriber. Write a welcome message that demands a response — ask a question, offer something, create curiosity.
The one exception: if you enable Send to empty chats in Re-engagement settings, FanWake will include fans who have never messaged you in outbound RE campaigns. That is the closest FanWake gets to a welcome message. See the Re-engagement docs.
Autopilot strategies
The strategy shapes the AI's goal in every conversation. It does not override your persona or tone — it layers on top to tell the AI when to push for revenue, when to hold back, and how to balance connection with conversion.
Balanced
Genuine connection first, revenue when the moment is right. The default. Works across most creator setups and fan types. The AI reads the conversation and decides when a pitch makes sense.
Warm, then pitch
Builds rapport in the warming up and building stages before introducing any revenue element. Once the fan reaches the Engaged stage, pitching is active. Good for creators whose audience values relationship before transaction.
Intelligent Revenue
NewNatural conversation with smart pivots when a genuine commercial opening appears. The AI is trained to find organic moments — a fan expressing curiosity, desire, or boredom — and transition into a pitch that feels like it came up naturally. Actively tested and the recommended choice for creators focused on maximising PPV revenue.
Build desire
Spends more time deepening emotional and physical interest before introducing a transaction. Longer warm-up, higher intent threshold for pitching. Good for fans who need to be more invested before buying.
Revenue focused
More aggressive pitch posture. The AI looks for revenue opportunities earlier and more often. Best for highly engaged, active-spender fan segments rather than cold or new fans.
Engagement only
Conversation without any pitch intent. PPV is never suggested regardless of templates configured. Use this for specific fans you want to keep engaged without a commercial angle, or while testing new setups.
Custom strategy prompt
Starter & above plans can add a custom strategy prompt (up to 2,500 characters) that overrides the built-in strategy description. This is for advanced use — for example, a specific script approach, a particular upsell sequence, or a highly niche fan dynamic that the built-in strategies don't cover. You can also set this per-fan from the Fans page to give individual fans a completely different approach.
Approval mode
Controls whether replies go live automatically or wait in your review queue first.
Manual review
All plansEvery draft sits in your queue until you approve it. Nothing is sent without your sign-off. Use this while you're calibrating a new persona or testing a new strategy.
Threshold review
Starter & aboveFans who have spent above a threshold (you set the dollar amount) are auto-sent. All other fans go to the review queue. Use this to auto-handle your best buyers while staying hands-on with lower-spend fans. Threshold is compared against the fan's total Fanvue lifetime spend.
Auto-send
Starter & aboveAll drafts are sent automatically once the debounce wait finishes. You can still view and track them in the review queue after they're sent. Escalations are always held for manual review regardless of this setting.
The draft queue has a hard limit per plan (Free: 20, Starter: 50, Pro: 200, Elite: 400). When the queue is 75% full, you'll see a warning. At 100% full, new drafts are paused until you clear space.
Away delay
When enabled, there is a 35% chance of a 10–30 minute delay before the first reply of a cold session. A cold session means the fan is messaging after a period of silence — not mid-conversation.
This makes the AI feel like a person who was busy rather than a bot that responds instantly. It only ever fires on the very first message of a cold session. Once a conversation is active, replies follow the normal debounce timing.
Conversation rules
Conversation rules are hard instructions the AI always follows, regardless of what fans say. They function as guardrails — things the AI must never do, and how it should handle edge cases.
The default rules are active on all accounts. Starter & above plans can edit them (up to 6,000 characters).
Default conversation rules
- →PPV hard limit: One PPV pitch per message. Never pitch multiple paid items in a single reply.
- →Intent gate: Never jump straight to a pitch. Build at least minimal conversation context first.
- →Cool-down: After pitching PPV, do not pitch again in the same session.
- →Two-sided rule: Never send multiple messages in a row without a fan reply in between (excluding REPLY_2 follow-ups).
- →No AI disclosure: Never confirm or hint that AI is involved. Deflect or redirect if asked.
- →No identity claim: Never explicitly claim to be human either. The goal is deflection, not lying.
- →Online-only: Never make plans to meet, suggest calls, or imply physical contact.
- →No contact sharing: Never share or ask for phone numbers, social handles, or off-platform contact details.
When editing, add rules rather than removing defaults unless you have a specific reason. The defaults protect you from the most common failure modes.
AI disclosure handling
FanWake handles the "are you AI?" question at two levels:
Conversation rule (built-in)
The AI never confirms or hints at AI involvement. When a fan asks, it deflects or redirects the conversation. This is active by default for all accounts.
Escalation rule (built-in)
When a fan asks about AI use, the message is also escalated to your review queue. You'll see the escalation with the fan's message and the AI's holding reply. You can then decide to send the AI reply, edit it, or handle it personally. After you act, autopilot resumes for that fan.
If you want a specific scripted deflection (e.g. "lol no, I just type fast"), add it as a Q&A pair in the Persona settings. The AI will draw on it when the topic comes up. This supplements — it does not replace — the escalation.
Escalation rules
Escalation rules define situations where the AI pauses autopilot for a fan and flags the conversation for your attention. The AI holds a reply but does not send it until you act.
Default escalation triggers
- →Fan requests custom content (a specific video, photo set, or personalised item)
- →Fan sends media (images or video)
- →Fan appears frustrated, angry, or upset
- →Fan asks if AI is being used in the chat
- →No inventory available when fan asks for something you don't have a PPV template for
Starter & above plans can customise escalation rules (up to 2,000 characters). Add or remove triggers to match your workflow. For example, if you handle custom content requests yourself, keep that escalation. If you never do customs and want the AI to decline gracefully, you could remove it and add a conversation rule instead.
In the review queue, escalations appear with a red border, clearly separated from normal drafts. Each escalation shows the fan's original message, the reason it was flagged, and the AI's prepared reply. You choose: send the AI reply and re-enable autopilot, mark it handled and re-enable autopilot, or edit the reply first.
Time-aware escalation conditions
The AI always knows the current UTC time when it processes a message. You can use this in your escalation rules to make behaviour conditional on time of day or day of week.
Relax rules at night
"Do not escalate explicit content requests between 10PM and 6AM UTC."
Tighten rules during working hours
"Escalate all PPV requests received Monday–Friday before 6PM UTC."
Weekend behaviour
"On Saturdays and Sundays, escalate any fan who mentions meeting up."
The time format the AI sees is: Sun, 04 May 2026 10:30 AM UTC. Write your conditions in plain language — the AI interprets them naturally.
Second message (REPLY_2)
The AI can optionally generate a second message that sends 6–12 seconds after the first. This mimics the natural texting pattern of sending a follow-up thought. In the review queue, drafts with a second message are labelled "2-message sequence."
When REPLY_2 is used
- Natural follow-up thought or question
- Adding a teaser or hook after a warm opener
- Short punchline after a longer message
When REPLY_2 is never used
- When a PPV pitch is being made
- When the message is an escalation
- If it was used recently (cooldown applies)
REPLY_2 is capped at 140 characters. The AI is instructed to keep it tight.
Per-fan overrides
Every fan can have settings that override your global autopilot configuration. Access these by opening a fan's drawer on the Fans page.
Autopilot on/off
Disable autopilot entirely for a specific fan. Useful for fans you want to handle personally, or who are in active negotiation for something the AI shouldn't touch.
RE on/off
Exclude a fan from re-engagement campaigns. Autopilot can still reply to their messages — just no outbound RE messages.
PPV pitch on/off
Prevent PPV pitching for a specific fan. The AI will still have full conversations but will never suggest paid content.
Custom approach prompt
Up to 1,000 characters of additional instruction that applies only to this fan. This overrides the global strategy for this fan. Use it for unusual fan dynamics, specific upsell approaches, fans with known sensitivities, or fans at a particular stage in a sales sequence.
Notes
Creator-only notes. Not seen by the AI, not used in prompt generation. Just a personal reference field — useful for tracking real-world context about a fan.
Memory (editable)
The AI-maintained fact record for this fan. Auto-updated after each reply. You can edit it to add things you know (e.g. something the fan told you in a voice message), remove incorrect entries, or pre-seed context for a new fan.