> For the complete documentation index, see [llms.txt](https://tembi.gitbook.io/tembi-knowledge-base/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://tembi.gitbook.io/tembi-knowledge-base/product-news/ai-assistant.md).
# AI Assistant
Tembi's **Assistant** lets you talk to the platform. Describe the kind of webshops you're looking for, ask a question about the data, or hand off a task — and the Assistant runs it for you in chat. No menus, no hunting for the right filter combination.
### What you can do
The Assistant is built around three jobs — the same three you'll see suggested when you open a new chat.
#### 1. Find what you're looking for
Describe what you need in plain language and let the Assistant turn it into filters. For example:
> *"Show me Finnish webshops on Shopify with revenue above €1M that aren't yet using Posti."*
The Assistant builds the filters, applies them to your view, and the matching webshops appear on the right. From there you can refine the prompt, tweak filters by hand, or move on to the next step.
If you've set up an **ICP** in Agent Studio, you can pin it at the top of the chat. Every prompt is then interpreted against that context — so you don't have to re-explain who your customers are every time. More on this below.
#### 2. Get answers about the platform
The Assistant is connected to this knowledge base, so it can answer questions about how Tembi works without sending you off to search for the right page. Ask it what a column means, how the Checkout Score is calculated, what counts as a "secondary export market", or how to set up an agent in Agent Studio — and you'll get a direct answer in the chat, with a link to the source article if you want to read more.
Because it pulls from the same articles you're reading now, answers stay in sync with the platform: as we add and update documentation, the Assistant gets smarter alongside it. If a topic isn't covered yet, it will tell you — and you can ping us at [**hello@tembi.io**](mailto:hello@tembi.io) so we can fill the gap.
#### 3. Take actions in the chat
Once you've found the right webshops, the Assistant can also act on them. For example, you can ask it to:
* **Save the current filter** so you can reuse it later
* **Assign tags** to a group of webshops at once
* **Assign a lead owner** to route accounts to a colleague
* **Set a lead status** in bulk
A prompt like *"Save this filter as 'FI Shopify mid-market' and assign all the webshops to Anna"* will do both in one go.
### Pair it with Agent Studio
The Assistant works best when it has context about your business. That's what **Agent Studio** is for. In Agent Studio you describe your **Ideal Customer Profile** and how you work with account management — the criteria that matter, the data points you weigh, the way you want leads to be handled.
Once your agent is set up, it shows up in the **ICP** dropdown at the top of the Assistant. Pick it, and every prompt from that point on is interpreted with that context. The Assistant can find webshops that fit your criteria, prioritise them the way you would, and run the right follow-up actions on your behalf.
→ [Agent Studio: Set up your agent](https://tembi.gitbook.io/tembi-knowledge-base/product-news/ai-agents/agent-studio-set-up-your-agent)
### Getting started
1. Click the **✨ Assistant** button in the top left of the Webshops view.
2. Pick a starting point — either type your own prompt or click one of the three suggestions: *Find leads*, *Get answers*, or *Perform actions*.
3. *(Optional but recommended)* Select an **ICP** at the top of the panel — set up in [Agent Studio](https://tembi.gitbook.io/tembi-knowledge-base/product-news/ai-agents/agent-studio-set-up-your-agent) — so your prompts are scoped automatically.
4. Use **Recent chats** to jump back into earlier conversations, or **New chat** to start fresh.
5. Click the expand icon to give the chat more room when you need it.
### Tips for good results
* **Be specific about the market.** *"Webshops in DK and SE"* gets you better results than *"Nordic webshops"*.
* **Mention the data points you care about.** Size, growth, tech stack, delivery providers — the more concrete the prompt, the tighter the filter.
* **Iterate.** It's a conversation. If the first result is too broad, say so: *"Now narrow it down to those without a 3PL system."*
* **Ask before acting.** If you're unsure what an action will do, ask the Assistant to describe it first before you run it.
### Beta notes
The Assistant is still being shaped, and it occasionally needs a second nudge to get a complex prompt right. If you hit something unexpected — or have ideas for prompts you wish it handled better — let us know at [**support@tembi.io**](mailto:hello@tembi.io).
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter:
```
GET https://tembi.gitbook.io/tembi-knowledge-base/product-news/ai-assistant.md?ask=
```
The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.