Can You Trust AI With Your Marketing Data? Scott on Smart Agency Masterclass

I joined Jason Swenk on the Smart Agency Masterclass to talk about something a lot of agencies are quietly getting wrong right now: trusting AI with their marketing data.

Let me be clear up front, because this is easy to misread. I am not anti-AI. Used well, it saves an enormous amount of time, and it can take your team well beyond just cranking out blog posts and ad copy faster. The problem is what happens when you point AI at attribution and performance data and assume the confident answer it gives you is the right one. Because AI is very good at sounding smart, and it is often wrong with total confidence.

The core issue is that large language models do not really understand time, causality, or intent unless you force them to. Marketing data is full of exactly those things. A conversion today might trace back to a first touch six weeks ago. A channel that looks like a winner might just be taking credit for demand that already existed. If you feed AI bad inputs or ask it the wrong question, you get very confident, very wrong advice, and then you make budget decisions on top of it.

The fix, and this is the heart of the conversation, is intention. Attribution only works when you define upfront what each campaign is actually meant to do, and therefore how it should be judged. Once intention is set, the data has a frame, and so does any AI you point at it. That is a big part of what the Five Forces framework is for: giving you a sane, rule-based way to read performance instead of reacting to whatever number looks good today.

Jason and I also get into why ROAS obsession leads agencies astray, especially when you are not separating new customers from repeat ones, and how Scale, Chill, and Kill zones take the emotion out of optimization decisions. If you run an agency and you are leaning on AI to interpret client data, this episode is worth your time before it costs you a budget or a client's trust.

Watch the full conversation below.