The tools I use in 2026

Two years ago I started testing AI tools on real client work. Not on demos or side projects, but on the briefs, scripts and campaigns that actually had to land. By the time I stopped counting, I'd spent €10,847 on subscriptions. Most of them didn't survive their first week.

That sounds like a bad investment. It was also the best education I've had in years. Once you've watched enough tools fail, you start to see why the few good ones work.

The tools that promised to think

I made the mistake almost everyone makes. I went looking for tools that would think strategically for me. Positioning, insight, a campaign idea in thirty seconds.

What came back sounded fine and meant very little. Generic, slightly artificial, and disconnected from the business it was supposed to serve. My clients noticed straight away. So did I.

It turns out I was in good company. Research from MIT published in 2025 found that about 95% of corporate generative AI pilots delivered no measurable return, and the researchers pointed at how companies used the tools rather than the quality of the models. That matches everything I've seen: AI is brilliant at execution and poor at judgment. The moment I stopped asking it to decide and started asking it to deliver, it became genuinely useful.

Thinking and writing

Claude and ChatGPT are where most of my days start. I give them the context a good colleague would need: the business, the positioning, the constraints, the goal. Then I ask for options, structure, variations or a second opinion on a script. They help me explore faster. They don't tell me where to go.

When ideas come faster than I can type, I talk instead. Wispr Flow turns spoken thoughts into clean text, which helps with a first draft, and even more when the tone needs to sound like a person rather than a template.

I've also started working with agents. OpenClaw, an open-source AI agent that runs on your own machine, can take a task and actually carry it out instead of only talking about it. It's still early, and it needs clear instructions and sensible limits, but it's the clearest glimpse I've had of where this is heading.

Making things

Video is my home turf, so this is where I'm most sceptical, and most impressed. Runway, Kling and Higgsfield won't replace a good shoot. They're excellent when an idea needs to be seen before anyone commits a budget to it, or when one strong piece of footage needs to become ten.

Once the creative direction is set, Canva lets teams produce on-brand assets without someone checking every pixel. It isn't a strategy tool. It's a scaling tool, and that difference matters more than most people think.

This website runs on Webflow, and YouTube Studio is where the video work lives once it's out in the world.

Systems that keep everything moving

This is where AI quietly saves me the most time. monday.com holds project structure, timelines and client visibility. Zapier connects the tools so information moves without anyone copying and pasting it. Calendly takes the back-and-forth out of booking a conversation.

None of it is glamorous. But good strategy dies without good systems, and I'd rather build a system once than run the same sprint every month.

Finding the right people

For growth, LinkedIn and Meta are where the audience is. Apollo and Clay help find and enrich the right contacts, so outreach starts with research instead of guesswork.

What I leave alone

I avoid “AI strategy” tools that promise insight without context, all-in-one platforms that do everything a little bit badly, and anything that takes creative control away from the people doing the work. If I can't explain to a client why we chose a direction, the tool isn't helping. It's hiding the problem.

Four questions every tool has to answer

Before anything new joins the stack, I ask:

1. Does it make the client's work better? Faster but worse is still worse.
2. Do I stay in control of the direction? If I can't steer it, it's automation, not strategy.
3. Does it buy me time to think? The goal isn't speed. It's more room for judgment.
4. Can I defend the output? If I can't explain the decision, the tool failed.

Fewer tools, used well

The list above looks long, but most of the real gains come from two or three tools I know inside out. Everyone has access to AI now. The advantage is knowing what to ask, what to ignore, and when to step in as a human.

AI didn't replace strategic creative work. If anything, it made it easier to see who was doing it in the first place.

If you're working out which of these belong in your own team, I'm happy to think it through with you. The full toolstack is on my About page, or you can book a 1-1.

— Joost