our

AI usage is always declared - internally, to our clients, and to our clients' clients. The use of AI tools is agreed upon at project start, with transparent model selection. We never use AI as a "secret ingredient."

If something goes wrong, we never blame the AI. As individuals and as an agency, we always take ultimate responsibility for the content we produce. We adopt an "expert in the loop" method: for every AI-enhanced process, there's a human setting the initial phase (objectives, constraints) and a human - with concrete expertise on the topic - evaluating the final output.

Before using any AI tool, we evaluate its privacy and security guarantees: for us, our clients, and their clients. We avoid sharing our clients' data and content with AI models we don't have full control over whenever possible. We assess potential risk on every project.

AI shouldn't become synonymous with low quality. When we use generative AI, we guarantee our clients quality equal to or better than what's achievable without AI. Where this isn't possible, we simply avoid using it. We constantly monitor outputs with a continuous improvement mindset.

AI shouldn't become synonymous with low quality. When we use generative AI, we guarantee our clients quality equal to or better than what's achievable without AI. Where this isn't possible, we simply avoid using it. We constantly monitor outputs with a continuous improvement mindset.

AI shouldn't be the door that lets bias into our communication. We carefully ensure that content created with generative AI is as free as possible from harmful and discriminatory bias. At the same time, we believe AI can help us make communication more inclusive and stereotype-free.