I build marketing and media programs that use AI to move faster and see further, grounded in the one thing no model can automate: understanding what actually makes a person stop, care, and buy.

I've spent more than 20 years in media and growth, from partner at an agency to senior leadership at a larger one, with CPG brands at the center of most of it. The technology has changed beyond recognition in that time. The buyer hasn't. People still decide with their gut and justify with their head. They still buy from brands that seem to get them.
AI makes me faster at everything around that moment. Pulling 31 months of data into a clear story. Spotting the broken pixel, the junk traffic source, the lead that never reached sales. Testing more ideas than a team could test by hand.
But the strategy starts with a human question, and it always will: what does this person want, what's in their way, and why should they believe us? Get that right and the machines amplify it. Get it wrong and they amplify that too.
Where the money goes, why, and what it has to return. Channel mix, flighting, and planning built around how real people actually use media, not how a platform wants you to buy it.
Google, Meta, email, marketplace. Account architecture that scales, testing that teaches, and reporting that tells you what to do next.
Most underperforming programs are not underperforming. They're mismeasured. I find the leak first, then spend.
Practical AI that makes teams faster and sharper, designed so human judgment stays in charge of the work.
Reported ROAS of 0.57. Actual blended ROAS of 1.15. Then we made it better. Read the case →
A senior living brand's best prospects were vanishing between the form and the CRM. Read the case →
Low-ticket marketplace sales were masking a core site that was starting to grow again. Read the case →
I design AI workflows for marketing teams around a deliberate rule: human point of view lives in its own space and never gets quietly blended into the machine's reasoning. The AI does the heavy lifting. People make the calls. That's not a limitation. It's the design.
Let's figure out what's in the way. Usually it's one of three things: the wrong audience, the wrong message, or numbers you can't trust.