Agent Templates
Agent templates are ready-made configurations for chatbots and AI agents: role, tone, greeting and handover rules are already written. At SendSeven they are 55 prompt presets in five categories and eleven languages that you load in the bot editor and adapt.
What are agent templates?
Agent templates are ready-made configurations for chatbots and AI agents. Instead of building a bot from scratch and writing the role, tone, rules and greeting yourself, you start from a template, adapt it to your brand and publish it.
The term is common in the industry and not defined consistently. Some providers mean finished bots, others only prompt building blocks. What they share: a template is a starting point, not a finished product.
What a template contains
- Role and job: what the bot is responsible for, and what it explicitly is not.
- Tone: how formal or relaxed the answers are.
- Greeting: the first message the contact sees.
- Handover rules: when a colleague takes over.
- Fallback: what the bot says when it cannot help and nobody is available.
Agent templates at SendSeven
SendSeven has two bot types, not a dozen template categories:
| Component | What it does | Typical use |
|---|---|---|
| AI assistant | Answers in natural language from your knowledge base (RAG), cites sources and hands over when confidence is low. | Support, sales, FAQ |
| Keyword bot | Matches incoming messages against a keyword list and replies with a fixed message, media, a template or a button menu. No AI. | Away message, simple menu |
| Flow | A component of its own, not a bot type: a sequence of steps such as send, wait, branch, collect input, tag, webhook. | Multi-step sequences, data capture |
The templates are not ready-made bots. They sit in the bot editor as prompt presets: 55 of them in five categories (support, sales, FAQ, email and a free one) across eleven languages. Each one describes the role, the tone, the wanted and unwanted behaviour and how the handover works. You load a preset into your AI assistant's prompt and keep editing it there.
From template to your own bot
- Pick the bot type: AI assistant for free-text questions, keyword bot for fixed keywords.
- Load a prompt preset: choose the category and language, then rewrite it in your brand voice.
- Connect the knowledge base: for the AI assistant, select the folders it may answer from.
- Set the handover: which confidence score makes a human take over, and which team that is.
- Publish: decide where the bot works, in the website widget, on a channel such as WhatsApp, Telegram, Messenger, Instagram or email, or through the API.
- Test: walk through it yourself once, including the cases the bot cannot handle.
Why start from a template
- Faster start: you do not write a prompt from zero, you shorten and sharpen an existing one.
- Completeness: the points beginners forget, such as fallback and handover, are already there.
- Multilingual: the presets come in eleven languages, not only English.
- Consistent quality: when several colleagues build bots, they do not sound completely different.
When a template is not worth it
- When your use case is so specific that no category comes close.
- When the bot mainly needs data from your own system. A flow with a webhook gets you further than a prompt.
- When your brand voice is very distinctive. The template is then only a skeleton you rewrite completely anyway.
For most small and mid-sized companies the route through a template is the shorter one. What an AI chatbot looks like in daily use is shown on our AI bots page.