Knowledge Base

Quickly find answers, guides and tips for AI Assistent. From first steps to advanced settings.

🔍

No articles found for your search query. Please try a different search term.

🚀

Getting started

🚀

Getting started with AI Assistent

Reading time: ~5 min · Beginners

▼

Welcome to AI Assistent

In this article we walk through the first steps to get your AI chatbot online quickly. Within 15 minutes you will have a working bot on your website.

Step 1: Create an account

Go to /register and create an account with your email address and company name. You will immediately receive a confirmation email. Click the confirmation link to activate your account.

💡 Tip: Choose a password of at least 12 characters. We recommend enabling two-factor authentication (2FA) immediately via Settings → Security.

Step 2: Configure your first bot

After logging in you will see the dashboard. Click “Create new bot”. Fill in:

Name: the name the bot uses to introduce itself to customers (e.g. “Emma” or “Assistant”).
Persona: a short description of the character and tone (e.g. friendly, professional, helpful).
Greeting: the first sentence the bot sends when a customer opens the chat window (e.g. “Hello! How can I help you?”).

Step 3: Copy and place the widget code

In the portal go to Settings → Widget. Copy the generated snippet:

<script src="https://interaip.ai/widget/{your-slug}/loader.js"></script>

Paste this snippet just before the </body> tag on every page of your website where you want to show the chat.

Step 4: Have your first test conversation

Open your website and click the chat button in the bottom right. Ask a test question. If the bot responds, the integration is working correctly. In the portal you will see the conversation appear immediately in the Conversations overview.

Tips for a good bot persona

Keep the name recognisable and fitting for your brand.
Describe the tone in a maximum of 2 sentences: the AI automatically adapts its style accordingly.
Add specific knowledge via the Knowledge Base or by having your website crawled — the more context, the better the answers.
Test regularly with questions your customers ask and adjust the knowledge base based on the results.

📦

Installing the widget on your website

Reading time: ~6 min · Technical

▼

Integrating the Widget

The AI Assistent widget is loaded via a single JavaScript file. The script is asynchronous and has no impact on the loading speed of your page.

HTML snippet

Add the following line to every page where you want to show the chat, just before the closing </body> tag. Replace {your-slug} with your own tenant slug (visible in the portal).

<script src="https://interaip.ai/widget/{your-slug}/loader.js" defer></script>

WordPress

You have two options:

Plugin (recommended): Install the AI Assistent WordPress plugin from the plugin directory. Enter your tenant slug in the plugin settings. The widget will be loaded automatically on all pages.
Manually via footer: Go to Appearance → Theme Editor → footer.php and paste the snippet just above <?php wp_footer(); ?>.

Shopify

Go to Online Store → Themes → Edit code.
Open the file theme.liquid.
Paste the snippet just before </body>.
Save and preview your store.

🛒 For Shopify Plus you can also add the snippet via a Custom Liquid section in the Theme Editor, so you don’t need to edit the code directly.

Other CMS systems (Joomla, Drupal, Squarespace, etc.)

Look in your CMS for the option to add custom scripts or custom code to the footer. Paste the snippet there. If your CMS supports a Custom HTML widget, that works too.

Customising the styling

Via the portal (Settings → Widget → Appearance) you can adjust:

Colour: choose the primary colour of the chat button and the chat window header.
Position: bottom right (default), bottom left.
Delay: how many seconds after the page loads before the chat button appears (0 = immediately).
Welcome message: an automatic proactive greeting after X seconds.

⚠️ Note: Only use the widget on domains you have added under Settings → Allowed domains. Requests from other domains will be blocked.

🤖

AI & LLM

🧠

LLM options explained

Reading time: ~5 min · General

▼

What is an LLM?

A Large Language Model (LLM) is the AI brain behind your chatbot. It processes your customers’ questions and generates appropriate answers based on the context and the knowledge base you have configured. The quality, speed and privacy characteristics of the answers are largely determined by the chosen LLM.

Available options

AI Assistent supports three types of LLM providers:

1. Platform LLM Recommended for starters

We manage our own LLM environment running on servers in the Netherlands. You don’t need to configure anything — the bot works immediately after activation. All data remains on European soil and you benefit from our management, updates and SLA.

2. Cloud AI: Claude & Gemini Fastest responses

You connect your own API key from Anthropic (Claude) or Google (Gemini). Responses are on average faster and the models are more up-to-date. Data is processed by the respective cloud provider; check their privacy policy for GDPR compliance in your situation.

3. Bring Your Own LLM (Ollama) Maximum control

You run your own Ollama server on your own infrastructure. No data leaves your network. Requires a Business or Enterprise plan and technical knowledge to set up.

When to choose which option?

Scenario	Recommended provider
Quick start, no technical knowledge	Platform LLM
Best quality & fastest responses	Cloud AI (Claude / Gemini)
Full data sovereignty, on-premise	Own LLM (Ollama)
Healthcare or legal sector with strict requirements	Platform LLM or Own LLM
Experimenting with open-source models	Own LLM (Ollama)

💡 You can switch provider at any time via Settings → AI Provider without having to reinstall the widget.

🔧

Set up your own LLM (Bring Your Own)

Leestijd: ~8 min · Technisch · Business / Enterprise

▼

Connecting Your Own Ollama Server

With the Bring Your Own LLM option you run your own language model on your own server or cloud environment. All conversation data stays within your network.

Requirements

Plan: Business or Enterprise (see pricing page).
Server: Linux or Windows server with at least 8 GB RAM (16 GB recommended for larger models).
Network: the server must be reachable via a public IP address or domain name with HTTPS (for production).

Step-by-step installation

Install OllamaGo to ollama.com and download the installer for your operating system. On Linux:
```
curl -fsSL https://ollama.com/install.sh | sh
```
Download a modelChoose a model and download it. We recommend Llama 3.2 for a good balance between quality and speed:
```
ollama pull llama3.2
```
For more models see ollama.com/library.
In the portal: set the AI ProviderLog in to the portal and go to Settings → AI Provider. Select “Own LLM”.
Enter the endpoint URLEnter the URL of your Ollama server. Ollama listens on port 11434 by default:
```
http://your-server.com:11434/v1
```
Or locally (for development):
```
http://localhost:11434/v1
```
Test the connectionClick the “Test connection” button in the portal. You will see a green tick if the connection is successful.
Configure routingYou can set your own LLM as the default for all bots, or choose a specific provider per bot via Bot → Settings → AI Provider.

Troubleshooting

Error message	Solution
`Connection refused`	Check that Ollama is running (`ollama serve`) and that the firewall allows port 11434.
`Model not found`	Make sure the model is downloaded: `ollama list` shows available models.
Slow responses (>10s)	Consider a smaller model (`llama3.2:3b`) or add more RAM to the server.
`SSL certificate error`	Use a valid SSL certificate (e.g. Let’s Encrypt) for production environments.

🔒

Privacy & Security

🛡

Privacy & GDPR compliance

Reading time: ~6 min · General

▼

How We Handle Your Data

At AI Assistent, protecting your data and your customers’ data is our top priority. We are registered with the Dutch Data Protection Authority (Autoriteit Persoonsgegevens) as data controller and act in full compliance with the General Data Protection Regulation (GDPR).

No training on customer data

We never use your customers’ conversations or data to train or improve our AI models. Your data belongs to you and is used exclusively to operate your chatbot.

Data location: the Netherlands

All customer data is stored on servers physically located in the Netherlands, in ISO 27001-certified data centres. No data is transferred outside the European Economic Area (EEA), unless you explicitly choose a Cloud AI provider (Claude, Gemini) using your own API key.

Encryption

At rest: AES-256 encryption for all stored data.
In transit: TLS 1.3 for all connections between your customers, the widget and our servers.
Database: sensitive fields (API keys, tokens) are stored encrypted using application-level encryption.

Data Processing Agreement (DPA)

For customers who process personal data of their end-customers via AI Assistent, we offer a standard Data Processing Agreement. Request it via our contact form or send an email to privacy@interip.nl. Enterprise customers can negotiate additional contractual safeguards.

Right to erasure: data export and deletion

If one of your customers submits a request to have their data deleted (right to erasure, GDPR Article 17), you can action this directly from the portal:

Go to Customers → [find customer].
Click “Export data” to download a complete overview.
Click “Delete account” to permanently erase all personal data.

Deletion is irreversible and will be completed within 72 hours including backup systems.

Strict Privacy Mode

With Strict Privacy Mode (enable via Settings → Privacy) conversations are not stored on our servers. The bot responds in real-time but you have no access to conversation history. Suitable for environments with the highest privacy requirements (healthcare, legal, financial).

⚠️ When Strict Privacy Mode is enabled, conversation analytics and reports are not available. The AI also does not learn from previous conversations within your tenant.

❓

Frequently asked questions

❓

Frequently asked questions

10 questions answered

▼

What does AI Assistent cost?

AI Assistent is available from €19.95 per month (Starter plan). All plans and prices can be found on the pricing page. There are no hidden costs and you can cancel at any time.

How many bots can I create?

The number of bots depends on your plan:

Starter: 1 bot
Professional: up to 5 bots
Business: up to 20 bots
Enterprise: unlimited

Which languages are supported?

AI Assistent supports more than 30 languages. The most commonly used are: Dutch (NL), English (EN), German (DE), French (FR), Turkish (TR), Spanish (ES) and Italian (IT). The AI automatically detects the customer's language and responds in the same language.

Can I upgrade my plan?

Yes, you can upgrade at any time via Settings → Subscription → Change plan. The new price takes effect immediately and the difference is calculated pro-rata on the next invoice.

What are credits?

Credits are the unit of measurement for AI usage in AI Assistent. Each plan includes a monthly credit budget; in the dashboard you can see your usage in real time. Technically, 1 credit represents a certain number of AI tokens (where 1 token ≈ 4 characters), but you do not need to worry about this — your usage is always displayed in credits.

How does the knowledge base work for my customers?

You build a knowledge base by:

Uploading documents: PDF, Word, TXT — the AI reads the content automatically.
Creating Q&A pairs: direct question-and-answer combinations for frequently asked questions.
Crawling your website: enter your domain name and the AI indexes your website.

The bot automatically uses this knowledge base to answer your customers' questions.

Can I connect multiple websites?

Yes. Via Settings → Allowed domains you can add multiple domains. The widget works on all added domains. Requests from non-allowed domains are automatically blocked for security.

Is an API available?

Yes, AI Assistent provides a REST API with API key authentication. This allows you to integrate the bot into your own applications, CRM or other systems. API keys are managed via Settings → API → Keys. Documentation is available in the portal.

How do I cancel my subscription?

You can cancel your subscription via Settings → Subscription → Cancel. Your account remains active until the end of the current billing period. You will receive a confirmation email. Would you prefer personal contact? Please reach out via our contact form.

What happens when my credits run out?

You automatically receive a warning email when you have used 80% of your monthly credit budget. At 100% usage the bot is temporarily paused until you purchase more credits or the next month begins. You can purchase extra credits directly via Billing → Buy credits — they are available immediately.

✨

Platform Features Technology & capabilities

AI Chatbots: what are they and how do they help your business?

Discover how AI chatbots automate customer service and provide 24/7 availability.

AI Credits: how does the credits system work and what does a conversation cost?

The Einstein AI platform uses a transparent credits system. Learn what a credit is, how many credits an average conversa

Knowledge storage & documents: how the AI learns from your business information

Upload your manuals, FAQs and product documentation and the AI learns from them instantly. Discover how the knowledge ba

Channels: connect your chatbot to website, WhatsApp, email and more

Your AI chatbot is not limited to one channel. Discover how to connect the bot to your website, WhatsApp, email, Telegra

Team members & roles: managing staff and setting permissions

Invite colleagues, assign roles and determine who can do what in the Einstein dashboard. From read-only access to full a

Custom AI personality: tailor the tone, style and knowledge of your bot

Give your AI chatbot a unique personality that fits your brand. Set the name, language use, conversation boundaries and

Einstein NL: AI on Dutch servers for maximum privacy and GDPR compliance

Einstein NL processes all AI conversations and user data exclusively on servers in the Netherlands. Ideal for organisati

Premium AI models: Claude, Gemini and GPT — when do you choose which model?

The Einstein AI platform supports the best commercial AI models. Learn the strengths of Claude, Gemini and GPT and which

Website Crawler: automatically scan your website for knowledge base content

The built-in web crawler scans your existing website and automatically converts it into searchable knowledge base conten

White label: your own branding on the AI chatbot

With the white-label option your chatbot appears entirely in your own house style: your logo, your colours, your domain

Bring Your Own Key (BYOK): use your own LLM provider account

With the BYOK (Bring Your Own Key) option you connect your own API key from OpenAI, Anthropic or Google. Your conversati

Intelligent AI routing: how the system automatically chooses the right model

Hybrid Routing analyses every question and automatically routes it to the most suitable AI model. Simple questions go to

Strict Privacy Mode: local processing without cloud data transfer

In Strict Privacy Mode all AI conversations and documents are processed exclusively on your own servers or the EU server

Didn't find your answer?

Our support team is happy to help. You can also start right away with a free trial.

Contact us Try for free →