Multi-bot conversations

Running several Aimogen Pro chatbot personas in one conversation, how they are addressed, and when it is worth the cost.

Several bots can take part in one conversation, each with its own persona, model and knowledge.

AI Chatbot › Multi-Bot Chat

Configuration#

Per placement:

[aimogen-chat-form additional_bots="researcher,editor,fact-checker"]

How it works#

Bots are addressed by name. The visitor can direct a question at one — "@researcher what does the data say?" — and the bots can address each other, which is how a multi-turn exchange between them happens.

Strip @BotName From AI Messages controls whether the addressing prefix is shown or hidden. Hiding it reads more naturally; showing it makes who-said-what unambiguous.

When it is worth it#

Genuinely useful:

• Different expertise. A technical bot and a commercial bot answering the parts of a question each is suited to.
• Review. One bot drafts, another critiques. The critique often catches real problems.
• Debate. Two bots argue opposite positions and the visitor sees both cases.

Usually not worth it:

• Bots that differ only in tone. The visitor gets the same answer twice.
• Anything where response time matters. Every bot that replies is another request, so replies take proportionally longer.
• Anything where cost matters. Three bots means roughly triple the token spend per exchange.

Designing a multi-bot setup#

Give each bot a distinct job in its Bot Context, and say so explicitly:

You are Researcher. You find and state facts, with sources where you have them.
You do not write marketing copy and you do not make recommendations.

Editor handles wording. Analyst handles recommendations. If a question is
theirs, say so briefly and stop.

Without that, bots duplicate each other and the conversation becomes noise.

Limit the number. Two or three is manageable. Five is chaos and five times the bill.

Decide who speaks first. Otherwise every bot answers the opening message.

Related#

• Chatbot personas
• Controlling cost