Work & thinkingExperimentBuilding
Building with AI
I built this portfolio with AI as a collaborator. I prioritized the budget-allocation example and geo-holdout simulation over more visual refinement, so people could explore the ideas for themselves.
- Project
- A portfolio where people can explore the ideas as well as read about them.
- Built
- A scrolling 3D golf ball, a budget-allocation example and a geo-holdout simulation.
- My responsibility
- Decide where the effort goes, what the examples demonstrate and what is ready to publish.
Put the effort into the tools.
I could have kept experimenting with the visuals. Instead, I wanted to spend the AI tokens on the substance of the site: the allocation example and the location-based measurement simulation.
A reader can move a budget slider and see how the modeled result changes. They can choose markets in a holdout simulation and explore the comparison. Those interactions give people something to examine alongside the explanation.
AI helped turn the design and code into a working site. The examples are illustrative; they aren’t forecasts or tools for making real media-buying decisions.
The checks are part of the work.
A working interface still needs a clear explanation of what it shows. I’m responsible for the claims, the assumptions behind the examples and whether the experience makes sense to someone using it.
Another automated check can help, but it can also miss an error. The decision about what evidence is enough and what is ready to publish stays with me.
Coursework and practice.
In 2024, I completed ELVTR’s AI in Marketing program. This site is one example of how I’m continuing to work with AI.
Recommendations