Chris Brandt, chief marketing officer at Chipotle, was interviewed for a study run with Google on how chief marketing officers act as change agents in adopting artificial intelligence. His position is that the technology is a means, not an agenda.
His assessment of the market is that most companies are still experimenting rather than working at scale, and he repeats an agency’s line that everybody is talking about it, few are really doing it, and those who are are not doing it well.
Key findings
- The most valuable uses he describes are operational rather than creative. Cameras in every restaurant can now report what percentage were fully staffed on the line at peak, which previously required pulling tape and going through files.
- The same cameras are being tested for auto replenishment, watching how full the bins of fresh food are and signaling the grill when an ingredient runs low.
- The clearest working case is hiring. A companion tool sets up appointments and screens resumes for restaurant managers, which matters at a company that has to hire twenty thousand seasonal staff with four thousand managers doing the hiring.
- He put a requirement into everyone’s objectives for 2025 to do something with the technology, however small, purely so that people start to understand it.
- The company deliberately did nothing with non-fungible tokens, and he uses that as the test case for how it approaches hype.
Starting from the problem, not the tool
About a year before the interview he asked every agency what they were doing with the technology. The smaller ones were using it for mundane tasks that used to take a long time, including sorting thousands of video files by typing in a word. The creative agencies were using it more, mainly for concepting and ideation.
The company has no dedicated practice for it. It works through its chief technology officer and large technology partners, and the approach is to bring problems rather than to look for places to apply a technology.
The same logic governs agency selection. He is not hiring an agency because it uses artificial intelligence, but because of what it brings in creative and ideas, and the questions he actually wants answered are cross media measurement, where his media is most effective, how to win in social, and the best return on ad spend.
Where it is already working
The operational examples come from owning every restaurant. With thirty-seven hundred restaurants in the United States, staffing could previously only be seen in aggregate, and now the same cameras that watch the line can be asked how many restaurants were properly staffed at peak.
Machine learning is also being used in accounting to process leases. The stated benefit is redeployment rather than reduction, moving people into other jobs because the task no longer takes as long.
The manager companion is the use he is most enthusiastic about. He describes the opportunity as taking mundane tasks, reporting, and prompts off the plate of someone trying to run a restaurant, and says that so far it is working pretty well.
The brand argument
Asked about the fear that brands will converge because everyone is using the same models, his answer is that advertising already looks the same, pointing to a parade of celebrities loosely affiliated with the brands they appear for.
His counter is that sameness creates its own incentive. If everybody looks the same there is a reason for brands to be different, and someone will always think a little differently and start the cycle again.
He grounds that in his own brand’s specifics: real cooking, unadulterated and unprocessed food, and advertising built from testimonials by actual employees rather than voiceover and an actor.
He is also skeptical of the promise most often made for the technology in marketing. One to one marketing has not replaced mass marketing, he argues, because it is expensive and because companies know less about the individual than they thought.
Change, and what he wants proved
His view of disruption is deliberately optimistic. Predictions that technology will eliminate everyone’s jobs have mostly produced different jobs instead, and he argues people are hard wired to resist change because in prehistoric terms change was dangerous.
He extends the same point to hiring at his own company. Turnover has been low, but when people do leave, replacements turn out to be great and bring their own ideas, which is his argument that change can be good.
The largest upside he names is outside marketing entirely: the volume of healthcare data available, and the possibility of finding trends and patterns in it that lead to medical treatments.
What he asks of the study is proof rather than prediction. He wants real cases where the technology made something meaningfully faster or better, or enabled something that would not have happened, and where it has become normal practice rather than a one-off experiment. His stated strategy is to be a fast follower.
Scope and limitations
This is one conversation of about half an hour, and the speaker is candid that his own uses are still being tested and that he is unsure whether some of them count as artificial intelligence at all. He also raises the definitional problem himself, asking whether the study distinguishes artificial intelligence from machine learning.
Source
This page summarises Vision Interview: Chris Brandt, CMO of Chipotle, by Mariana Peneva, Chris Brandt, February 19, 2025.