AI Chatbot - Level.pro

Native AI Chat for CMS-On-Brand, No SaaS Routine query deflection ↑55% and p95 response time ↓60%. Native, secure AI chat for TYPO3 CMS: install, train, and launch in minutes-right from your backend. Key Results Deflection of routine queries: 55% Answer latency (p95): ↓ 60% after launch CSAT on assisted sessions: 4.8/5 Content freshness to index: < 15 minutes via scheduled crawls

Client

Level.pro

Industries

CMS / MarTech, E-commerce, SMB

Services

RAG Chat, LLM, Widget UX, Analytics

Project Scope

AI-Powered CMS-Native Chatbot Implementation

AI ChatCitationsMulti-botOn-brand

Challenge

Support queues were piling up with repeat questions spread across marketing pages, documentation, and product content. Existing chat widgets couldn’t deliver trustworthy, cited answers, didn’t match the brand, and lacked any real governance.

The team needed an on-site assistant that could reliably cite from current website content and files, respect access permissions and editorial workflows, and launch quickly without the hassle of a new SaaS contract. It also had to provide clear usage insights to show what visitors were asking and where content needed improvement.

Solution

Install → Train → Launch

Set up through the Extension Manager, connect your own AI key, and build a knowledge base from crawled pages, uploaded files, and custom Q&A. Style the widget to fit your brand and go live instantly.

Native to CMS

Manage multiple chatbots directly inside the CMS backend, assign them to specific pages or sites, and iterate seamlessly with editors, no external tools needed.

Grounded & Fresh

Domain-aware crawling, metadata mapping, and scheduled refresh keep responses accurate. Answers include cited passages to avoid guesswork.

Secure by Design

Your AI key, your data. Vectors are stored locally with full site-level governance, ensuring everything stays within your infrastructure.

Insights that Matte

Access chat history and query trends to refine content, optimize flows, and measure assistant performance over time.

AI Architecture Flow

AI Technical
Stack

Vector Store

RAG

OpenAI

Claude

Google Gemini

Azure OpenAI

Custom LLM

Frontend Widget

Backend Bot Manager

Session Logs

Query Insights

0%

Routine question deflection increased, freeing support teams to focus on complex issues.

0%

p95 response times improved, boosting engagement and conversions on key pages.

0+

A customizable widget and CMS-native workflows simplified rollout and ownership.

0+

Chat insights fed into content updates, reducing friction over time.

AI Architecture Flow

AI Chatbot

Client's
feedback

“T3AC let us launch an on-brand chatbot without another SaaS. It cites our own pages, stays fresh after each deploy, and cut our repetitive tickets in half.”

Susann T.

VP Marketing Level.pro GmbH

image (7)Your Next Project is a Conversation Away!

CTA ImageCTA Image
AI Chatbot: Faster Answers, Fewer Tickets