Dennis van de Velde

I'm Dennis van de Velde, SEO specialist at Space'M Online and founder of LearnLLM. I work with AI language models every day and write here about what works, what goes wrong and which checks are needed before output goes out under your name.

View the online AI courses Follow me on LinkedIn

SEO specialist at Space'M Online · Founder of LearnLLM · Previously at Rabobank, ABN Amro and Van Lanschot

Dennis van de Velde, founder of LearnLLM

Work experience of Dennis van de Velde

My work experience lies in SEO and in banking, and the two come together in how I work with AI.

Since March 2026 I have been an SEO specialist at Space'M Online in Eindhoven, the Netherlands. Before that I did the same work at Websitepromotor, also an SEO agency. At both agencies I use AI in almost every step of the SEO process, on client websites.

Before that I worked at Rabobank, ABN Amro and Van Lanschot, including in the fraud department. Nothing went out there without a second check. I now apply that four-eyes principle in all my AI workflows. From that period I also took away that you say honestly when you don't know something.

A language model can sound convincing without being right. That is why the check I apply depends on the risk: an internal note calls for something different from a published article or advice to a client.

In April 2026 I started LearnLLM, my own platform for working with AI in a practical and responsible way. What the platform offers and why it was set up is on the page about LearnLLM. Under my own name I also work as an SEO specialist in Eindhoven for local businesses.

My experience per AI tool

My experience differs per AI tool, so you know what my articles are based on.

Claude

Daily

Claude is my main work tool, in the interface and through the API. In the interface I write and analyse; through the API, tools run on my own server, such as analyses of Search Console data. Claude reads that data directly from BigQuery.

That way of working is also in the Claude course, with prompts and checkpoints for each task.

Automations

In production

I build my automations with AI as a building partner, with security first: which data goes in, where it ends up and who can access it. I update the lessons of my e-learnings with a script that generates the changes and checks them in advance.

How to set up such a connection yourself, without code through Zapier or Make, is covered in the AI Automation course.

ChatGPT

Used intensively, now switched

I worked intensively with ChatGPT for a long time and built several custom GPTs for keyword research, content briefs and technical SEO analysis. Based on my own comparison, I have switched to Claude for most of my work.

Where the two differ in practice is described in the comparison Claude vs ChatGPT.

Gemini

Limited, within Google Workspace

I use Gemini within Google Workspace, to summarise emails and spreadsheets and for analyses in Sheets. Statements about Gemini beyond those tasks are based on the documentation, not on my own experience.

Other AI tools

Tested

I have tested Perplexity and Grok to place them in the AI landscape. I write about these tools as someone comparing them, not as a daily user.

Keyword research

Deliberately by hand

I do keyword research by hand, based on data from a keyword tool. A language model does not know what people actually search for and sometimes gives search volumes it cannot know. Clustering and organising is something the model may do.

How I check AI output

I check AI output with fixed steps, and how thorough they are depends on the task.

A test on a small sample
Before I start a bulk job, I first run a test on a few items. That catches most errors before they repeat themselves a hundred times.
Automatic check on structure
Output that a script processes further is checked automatically for format and structure. If the format is wrong, processing stops.
Reading it myself before publication
I write my texts with Claude and read them myself before they go online. I tried a fully automated writing pipeline and dropped it again because the quality fell short.
Keeping prompts and output
I log prompts and output, so I can trace later how a result came about. That makes an error easy to find and fix.

What I build at LearnLLM

At LearnLLM I build and maintain the e-learnings and the knowledge base.

The e-learnings are for professionals who already use AI but do not yet have a fixed way of working. If you want to show that employees use AI responsibly, start with the AI Literacy course. If you mainly work with one tool, the ChatGPT course or the Claude course is a better fit.

Teams get an account and a certificate in their own name for each employee. Which format suits an organisation is explained on AI training for teams.

I also write and maintain the AI knowledge base, with more than a hundred articles on AI models, prompting and AI at work. When a feature, price or rule changes, I update the article.

Topics I write about

The topics I write about are spread across ten sections of the knowledge base.

All articles by Dennis van de Velde

All articles by Dennis van de Velde on LearnLLM, newest first.

Dennis van de Velde

Work experience of Dennis van de Velde

My experience per AI tool

Claude

Automations

ChatGPT

Gemini

Other AI tools

Keyword research

How I check AI output

A test on a small sample

Automatic check on structure

Reading it myself before publication

Keeping prompts and output

What I build at LearnLLM

Topics I write about

AI basics

AI models

Prompt engineering

Comparisons

Claude guide

ChatGPT guide

Gemini guide

Automation

AI use cases

AI for business

All articles by Dennis van de Velde

What is Google Veo? How to make videos with Veo 3.1

What is n8n? How it works and what you can automate

How to change Gemini language (and get answers in Dutch)

Grok pricing 2026: Free, Plus, Premium, Super and Ultra

What is Perplexity AI? How it works, models and your data

How to turn off Gemini: 6 ways on Android and Samsung

Grok models: every version from Grok 3 to Grok 4.7

Perplexity Pro price: what Pro, Max and Enterprise cost

AI agents for business: uses per department and sector

What is human in the loop? Why people stay responsible for AI

Zero-shot prompting: what it is and when to use it

What is a system prompt? How it differs from your own prompt

Is AI free? What free AI tools offer and when to pay

What is a context window? Why AI forgets earlier messages

Prompt injection: what it is, examples and how to limit the risk

What is RAG in AI? How retrieval-augmented generation works

AI writing: how to get usable text from a language model

Audio to text AI: how to transcribe recordings reliably

AI summarizer: how to summarise documents with AI

What is a neural network? Layers, weights and training explained

What is deep learning? How it differs from machine learning

What is an API? How requests, keys and integrations work

What is a webhook? How to use one in an automation

What is machine learning? Meaning, types and examples

What are tokens in AI and how do you count them?

Claude API key: how to create one in the Claude Console

Claude Pro usage limits: how many tokens do you get?

Gemini pricing: every plan and price in 2026

What is Power Automate? Flows, connectors and AI explained

What is MCP? The Model Context Protocol explained

Claude API: what it is and what you can do with it

What are AI agents and what can they do for your office?

AI for architects: where it helps and what to protect

AI process optimization: start with the bottleneck, not the tool

AI note taker: how to turn a meeting into minutes with AI

What is an LLM? Large language models explained

When did ChatGPT come out? Release date and timeline

Few-shot prompting: how examples steer AI output

AI in sales: where it works and where it clashes

AI in HR: what is allowed and what works

DeepSeek vs ChatGPT: a full comparison for 2026

AI for customer service: what works in a small team

What does GPT stand for? Generative pre-trained transformer

ChatGPT translation: how good is it really?

What is OpenAI? The company behind ChatGPT explained

What is Muse Spark? Meta’s AI model explained

ChatGPT prompts for work: 6 ready-to-use examples

ChatGPT Go vs Plus: which plan is right for you?

ChatGPT photo editing: what to expect and what to check

How to cancel your ChatGPT subscription (web, iPhone, Android)

ChatGPT for business: plans, data and the AI Act

AI productivity: six ways to work smarter with ChatGPT

Is ChatGPT safe? Privacy and reliability explained