The term AI native may soon start replacing digital native for marketing and tech companies. A new digital divide is forming, one that separates those who embrace AI tools and use it for everything and those who don’t.

People who understand generative AI tools have integrated it fully in their lives, for work and personal use. They don’t need to think about when and why to use genAI systems based on large language models. It’s obvious to them that these tools have already reshaped human existence and they are on board. They’re all in.

This always on genAI mindset is what future-fit companies increasingly look for in employees and partners. They are not navel gazing ad infinitum for charlatan consultant-doublespeak “use cases” to justify laser focusing on these systems as essential work tools. For them, the use case is revealed through using genAI habitually.

When ChatGPT launched three years ago, The Toast were early adopters as a professional business tool on our journey to transitioning to an AI native agency. We’ve expanded our paid tools to four top tier genAI platforms and regularly mix and match these tools in a constantly adaptable mix of engines to power everything from deep research, analysis, reporting, vibe coding and strategy to error correction and beyond.

As par for the course our tools have built up a huge knowledge base about how we operate, our agency style and processes, assisting us in speedily executing complex tasks with clarity.

As with everything in marketing, the magic is enabled by the brief, so we spend significant resources on creating, developing, refining and checking our briefs so that eventual genAI outcomes are high on a trust scale. Not that we abdicate anything. Far from it. We make sure genAI mistakes and off-brand generations are corrected in a learnable way, so that these occurrences are enhanced with focused guidance for future tasks.

Our agency has recently conducted proprietary genAI tool usage research amongst tightly defined audience groups. While it’s not unexpected to note that 90%+ of respondents use these tools frequently and that 81% use ChatGPT specifically, it’s noteworthy that Google’s Gemini app is rising in popularity, specifically since the launch of the excellent Gemini 3 Pro, as are Meta’s genAI helpers.

GenAI agents have so far underperformed for us, despite all the associated hype. In a recent conversation with a senior media business leader, he spoke about genAI agents as if his company was using them successfully. However, it took just one or two questions, to realise he was just BS’ing. This example could highlight one of the most important attitude lessons about genAI. Don’t believe any of the hype. Believe the results.

Not that we haven’t used agents. The results have simply been underwhelming. We’ll continue to watch what transpires carefully and let the agents loose when we reckon there is true value to be had. It’s a wait and see for now.

We have successfully fully integrated data repositories such as mail, storage and the like, to relatively speaking positive effect. The core tools are strong when it comes to contextualised searching, finding and analysing quickly, so integration has thus far yielded more useful results than agents.

Our best way of using genAI is mainly manual. We do the core thinking, develop deep problem statements, curate what we want from the tools and then check, correct and edit all results.

The pure act of having integrated genAI doesn’t mean we simply take it as it comes. Rather, we’ve become adept at productively using the tools for what they’re (currently) best at and we keep a keen eye on them as they develop. In the process we’re acquiring a useful amount of knowledge about how to best implement these tools as time goes by.

And in our book, learning offers a professional use case.