Romain Guery-OdelinPersonal · Ongoing · Always
Watching and testing AI tools
I never recommend an AI tool I haven't tested myself. This habit builds reusable building blocks: connectors, templates, generation pipelines.
Context
Personal, ongoing project: before proposing a tool or method to a client or a team, I test it on my own needs first.
The problem
Recommending an AI tool you haven’t tested yourself means risking selling a marketing promise. And rebuilding a technical building block from scratch for every new project wastes time.
What I did
- Canva connector: automatic generation of visuals from a template, tested and then reused to produce content at scale.
- WhatsApp connector: automated retrieval of information (results, events) as a trigger for a generation chain.
- Semi-automated communication pipeline: a template created once, recurring generation driven by a simple workflow, information pulled via WhatsApp, applied to the template, then human review before publishing on social media.
- Motion design and video generation: tools tested and compared to produce presentation videos without going through an outside vendor.
The result
These building blocks don’t stay isolated tests: the Canva connector and the communication pipeline power the sports club’s automated communications, while the WhatsApp connector and video generation serve other personal and professional projects. Watching tools only pays off if it turns into reusable building blocks — that’s what I check every time.
3 building blocks built through ongoing testing — a Canva connector, a WhatsApp connector, a semi-automated communication pipeline — reused across several other projects.