AI Vocabulary

AI Glossary

Understand the key AI concepts and how we apply them to solve real business problems for European SMEs. No jargon — just clear explanations with practical examples.

AI Agent

Core Concepts

An autonomous software program that can perceive its environment, make decisions, and take actions to achieve specific goals — without constant human supervision.

How We Apply It

We build custom AI agents that handle your customer service, document processing, scheduling, and data entry. Each agent is trained on your business data and operates within your security policies.

Real Example

A customer service agent that answers 70%+ of support tickets in 4 languages, escalating complex cases to your team.

Virtual Worker

Core Concepts

An AI-powered digital employee that performs specific business tasks end-to-end, mimicking the workflow of a human worker but operating 24/7 without breaks.

How We Apply It

We create virtual workers for repetitive, high-volume tasks like invoice processing, email triage, appointment scheduling, and report generation — freeing your team for strategic work.

Real Example

A virtual accounts payable worker that processes 500+ invoices/month, matching POs, flagging discrepancies, and routing for approval.

RAG (Retrieval-Augmented Generation)

Technical

A technique that enhances AI responses by first retrieving relevant information from your company's documents, then using that context to generate accurate, grounded answers.

How We Apply It

We build private RAG systems connected to your file servers, SharePoint, or databases. Your employees can ask questions in natural language and get answers sourced from your actual documents — with citations.

Real Example

An employee asks 'What's our return policy for B2B clients in Germany?' and gets the exact answer from your policy documents, with a link to the source.

Agentic AI

Core Concepts

AI systems that can plan multi-step tasks, use tools, and adapt their approach based on intermediate results — going beyond simple question-answering to actually completing complex workflows.

How We Apply It

We design agentic workflows where AI can research, draft, review, and execute multi-step business processes. The AI breaks down complex tasks, uses multiple tools, and self-corrects.

Real Example

An agentic system that receives a new client inquiry, researches the company, drafts a personalized proposal, schedules a follow-up, and updates your CRM — all autonomously.

LLM (Large Language Model)

Technical

A type of AI model trained on vast amounts of text data that can understand and generate human-like text. Examples include GPT-4, Mistral, Claude, and LLaMA.

How We Apply It

We select the right LLM for each use case — prioritizing European models like Mistral for data sovereignty, or using specialized models for specific languages or industries. All models run on EU-hosted infrastructure.

Real Example

Using Mistral Large for a French legal document assistant, ensuring all data stays within EU borders.

Fine-Tuning

Technical

The process of further training a pre-trained AI model on your specific data to improve its performance for your particular use case, industry, or language.

How We Apply It

We fine-tune models on your company's terminology, writing style, and domain knowledge — making the AI sound like a member of your team, not a generic chatbot.

Real Example

Fine-tuning a model on 5 years of your customer support conversations so it matches your tone and knows your product catalog.

Vector Database

Technical

A specialized database that stores information as mathematical representations (vectors), enabling fast semantic search — finding information by meaning, not just keywords.

How We Apply It

We use vector databases to power our RAG systems, enabling your AI assistants to find relevant information even when the exact keywords don't match. This is what makes AI search feel 'intelligent'.

Real Example

Searching for 'employee vacation policy' also finds documents titled 'PTO guidelines' and 'annual leave procedures'.

Prompt Engineering

Technical

The art and science of crafting instructions (prompts) that guide AI models to produce the desired output — including system prompts, few-shot examples, and chain-of-thought reasoning.

How We Apply It

Every AI agent we build includes carefully engineered prompts that define its personality, knowledge boundaries, safety guardrails, and output format. This is a core part of our development process.

Real Example

A system prompt that tells a customer service agent: 'You are a helpful assistant for [Company]. Never discuss competitors. Always offer to escalate to a human for complaints.'

GDPR Compliance

Compliance

The General Data Protection Regulation — the EU's comprehensive data privacy law that governs how personal data is collected, processed, and stored.

How We Apply It

Every solution we build is GDPR-compliant by design. We use EU-hosted infrastructure, implement data minimization, provide audit trails, and ensure the right to erasure is technically possible.

Real Example

An AI assistant that automatically redacts personal data from training datasets and logs all data access for audit purposes.

AI Act

Compliance

The EU's regulatory framework for artificial intelligence, classifying AI systems by risk level and imposing requirements for transparency, safety, and human oversight.

How We Apply It

We design all solutions with AI Act compliance in mind — documenting risk assessments, ensuring human-in-the-loop for high-risk applications, and maintaining transparency about AI decision-making.

Real Example

For an HR screening tool (high-risk under AI Act), we implement mandatory human review, bias testing, and detailed documentation of the AI's decision criteria.

Workflow Automation

Core Concepts

Using software to automate repetitive business processes — from simple if-then rules to complex multi-step workflows involving multiple systems and decision points.

How We Apply It

We use tools like n8n and custom integrations to connect your existing systems (CRM, ERP, email, etc.) and automate data flows, notifications, approvals, and reporting.

Real Example

When a new order arrives in Odoo, automatically generate an invoice, notify the warehouse, update the CRM, and send a confirmation email to the customer.

Data Pipeline

Technical

An automated system that extracts data from various sources, transforms it into a usable format, and loads it into a destination for analysis or AI training.

How We Apply It

We build data pipelines that clean, structure, and prepare your business data for AI consumption — connecting disparate systems and creating a unified data layer.

Real Example

A pipeline that combines sales data from Odoo, customer feedback from email, and market data from external APIs into a unified dashboard with AI-powered insights.

Have questions about any of these concepts?

Get My Free AI Assessment