AI Employee
Your own AI employee.
An AI system that knows you, uses your tools and remembers your company's history. This isn't the bot for your customers on your website — this is your own helper.
Difference from the Website Bot
Owner-facing, not customer-facing.
The Website Bot talks to your customers on your site. The AI Employee works for you as the owner: triage email, remember notes, work in your systems, do research, be there every day. Two different products for two different jobs.
AI Employee
For you as the owner
Email triage, note search, CRM upkeep, reports, research
What it can doWhat makes the difference
It remembers. Permanently.
Other AI tools forget everything after each chat. Your AI Employee builds a private knowledge base over months: your customers, your processes, your decisions. What you decide today, it still knows tomorrow.
Persistent memory
Every interaction flows into a knowledge graph with entities, relationships and observations. No throwaway chat threads, long-term company knowledge.
Your data, your storage
The memory lives in your database, not in someone else's cloud. You can export, archive or delete it any time.
Bi-temporal queries
Ask it *What was the status on March 15?* and it pulls the knowledge snapshot from the past. Audit-ready.
Cross-tool compatible
The same memory works in Claude Desktop, ChatGPT, Cursor and Codex. You switch the tool, not the memory.
What it can do
Concrete tasks instead of vague promises.
An AI Employee without clear tasks is just a toy. Here are eight tasks it takes over from day one:
Email triage
Distinguishes important from newsletter, drafts replies, spots leads, archives spam. Uses your past replies as a style template.
Note research
You ask *What did we discuss with customer X last year?* and it finds it across email, notes and CRM entries — even when the wording is different.
CRM upkeep
Logs contacts, deals and interactions automatically. Links people to companies. Keeps the pipeline status current.
Report generation
Weekly status reports, monthly revenue overview, quarterly comparisons. From your real data, not from templates.
Meeting prep
Before a customer call it shows you the last conversation, the latest status, open items. You walk in informed.
Web research
Researches competitors, market news, sources. Stores results as a knowledge entry so you can find them later.
Document search
Finds PDFs, contracts, invoices via semantic search. Even when you don't remember the exact file name.
Tool execution
Connects to your tools — Stripe, n8n, Google, Notion, Slack — and runs tasks there. By instruction, not by click.
Who this is for
Four concrete SMB situations.
Solo consultant with too much email
Meetings all day, email pile in the evening. The AI Employee triages in the morning, drafts replies, keeps the CRM current. Saves 5-8 hours per week.
5-person real estate office
Inquiries come in via portals, email and WhatsApp. AI Employee gathers everything in one inbox stream, spots duplicate inquiries, prepares exposé responses.
Restaurant owner running daily ops
Bookings, supplier emails, staff questions — all in parallel. AI Employee handles standard questions, learns your reply patterns, only escalates what matters.
Agency selling AI services
You need an internal sparring partner that knows your projects, your code standards, your customer history. The AI Employee becomes the shared knowledge layer.
Variants
Three stacks. You pick by data sovereignty and budget.
The AI Employee comes in several variants. Which one fits depends on how sensitive your data is, how fast it needs to respond and how much setup budget you have.
AI Server
LiveFull data sovereignty, dedicated server
Your own Hetzner server in Germany with your AI system. Memory, tools and API all run on your side. You control every request.
from 299 €/mo
Hosting + maintenance. Setup billed by scope.
Cloud Stack
PlannedQuick entry without a server
Cloud-based variant via our managed memory SaaS. Lowest entry barrier, no server setup, but data flows through external LLM APIs.
from 99 €/mo
Preorder available.
Hybrid Stack
PlannedCompliance + cloud performance
Memory and data stay on your side, LLM calls go through EU-hosted cloud APIs. Compliance-friendly for sensitive sectors like law, consulting, healthcare.
from 199 €/mo
Preorder available.
Pricing
Range by stack and tool depth.
The AI Employee is a system of memory plus tools plus connection to your data. Price depends on which stack, how many tools, which data sources. We show you a clear range before setup.
Entry
99 - 299 €/mo
Cloud Stack with 3-5 tools, no dedicated server.
Standard
299 - 599 €/mo
AI Server with 8-12 tools, standard CRM and mail integration.
Custom
from 599 €/mo
Own integrations, multiple data sources, custom tools.
One-time setup fee between 990 and 4,900 €. Concrete quote after a 30-minute intro call.
FAQ
Common questions
What separates the AI Employee from the Website Bot?
The Website Bot is customer-facing — it talks to your customers on your website. The AI Employee is owner-facing — it works for you as the owner. Two different products. You can have both, but they do different things.
What separates this from ChatGPT with Custom GPTs?
Custom GPTs forget everything after each chat. The AI Employee has a persistent memory that grows over months. Plus it has real tool access to your data, not just API snippets. Plus your data stays with you, not with OpenAI.
Do I need my own server?
No. With AI Server we provide the server (Hetzner Germany). With Cloud Stack everything runs in our managed memory SaaS. You only need a browser or an MCP-capable AI client.
Which AI models does the AI Employee use?
Anthropic Claude (Sonnet 4.6 / Opus 4.7) as default, alternatively OpenAI GPT-5, Gemini 3 Pro or local models (Qwen 3, Llama 3.3). You pick by budget and privacy requirements.
What does it really cost?
Range 99 - 599 €/month plus a one-time setup between 990 and 4,900 €. Concrete number after a 30-minute call, because it depends on which tools, which stack and which data integrations you need.
GDPR and privacy?
AI Server in Germany (Hetzner Frankfurt or Falkenstein, EU). Standardised data-processing agreement. With Cloud Stack the LLM calls run via EU servers of the chosen provider. With Hybrid Stack your data stays entirely with you, only LLM prompts leave.
How long does setup take?
Cloud Stack: 1-2 weeks from kick-off to live. AI Server with standard tools: 3-4 weeks. Custom integrations 4-12 weeks depending on depth. We define clear milestones upfront.
Can I keep building on it myself?
Yes. Code, memory schema and tools live on your side. We use open standards (MCP, OpenAPI, A2A). No lock-in. We can keep accompanying you or you continue alone.
Next step
30 minutes intro call.
We look at which stack fits your setup and which two or three tools give the biggest leverage. No sales pressure.