Discovery, research, and decision do not happen in the same place. Only on your dashboard.

By Danielle Sene, Senior Marketing Manager at Squadra

I'll start with the annoyance, because that's where the text came from.

We spent about ten years building infrastructure to answer a poorly formulated question. CDP, data lake, CRM plugged into media, database enrichment, lead scoring, propensity model, governance committee. All of this to be able to say, with some accuracy, who that person was: Millennial, Gen Z, Alpha. 

Then came microgenerations, behavioral segments, personas with stock photo names and faces (and yes, I've presented one of those more than once). Always with the same phrase on the last slide: the right message, for the right person, at the right time.

But people don't behave like personas. They never did. What changed is that today there's enough data to prove it, and we keep buying media as if there wasn't.

The State of the Consumer 2026, published by McKinsey in June, shows Gen Z, the generation we've conventionally called the most digital in history, discovering brands in physical stores in 28% of cases, compared to 23% on social media. Friends and family, 18%.

The lazy interpretation here would be to say that Gen Z is returning to offline. It's not, and the report itself dismantles this a few lines later, when it shows social media weighing more in the decision-making process than in the discovery process. This is precisely where it gets interesting. Discovery, research, and decision don't live in the same place, don't follow the same order, and don't have the same owner on your spreadsheet. We buy media as if they did.

And channel is exactly the unit by which we organize budget, goals, dashboards, teams, and board presentations.

Consumers don't organize their decisions by channel. They organize by context. They discover on Instagram, ask in a WhatsApp group, search on Google, watch a video from someone they don't even know, enter a store, and decide. They might do this in an hour or over three weeks. We call it a journey. For them, it's just Tuesday.

Now add AI to this picture.

Meta has generous numbers on this. A study it commissioned and released in June indicates that 84% of the Gen Z surveyed discover products and brands on its platforms, and that Reels accounts for 81% of discovery, 66% of consideration, and 47% of purchase influence. This is a vendor study measuring its own funnel in a single market, and even so, I believe it. The point is different: what it measures is influence. Influence and trust are not the same thing, and none of these numbers can tell the difference.

Yes, this is an argument about machine limitations. I work in technology marketing, so I'll do my best not to pretend this is nostalgia. AI will do what it already does very well: test faster, generate more variations, segment better, shorten the distance between an idea and a thousand versions of it. What it doesn't answer is the question that comes before: Does this message deserve attention?

This is where the math doesn't add up. We're getting very good at distributing messages at scale while the person on the other end becomes increasingly selective about which messages to believe. Two curves rising in opposite directions, and only one of them appears in the report.

In B2B, especially for those delivering technology, the situation is even more blatant.

Gartner surveyed 645 B2B buyers between August and September last year. Two-thirds (2/3) prefer to buy without talking to a salesperson. 45% used generative AI in their last purchase, mostly to gather information about suppliers and products. And 69% turned to a salesperson to validate what the AI had said.

Notice that these three things refer to the same buyer. They don't move from digital to human like someone going down a funnel. They keep both open and use each for a different purpose. Machine for gathering inventory, human for sharing risk. There were seven sources of information, on average, in a single purchase.

Because when the decision involves budget, downtime risk, the reputation of those who signed off, and an entire operation depending on it, they will want to talk to a client who has gone through the same project. They will want to look the expert in the eye. They will want to know who answers the phone when the schedule slips. No CIO approved a seven-figure contract because a dynamic creative caught their eye at the right moment of their scroll.

And Gartner measures the cost of ignoring this: purchases made through pure self-service generate much more regret afterward. The house projection is that by 2030, three-quarters of B2B buyers will prefer experiences that prioritize human interaction over AI. Not out of nostalgia. Due to accumulated poorly made decisions made alone.

This also applies to media. If anyone can produce three hundred creatives a day, production has ceased to be a competitive advantage. The bottleneck has returned to the idea and, before the idea, to understanding the context in which it needs to appear. A brilliant creative at the wrong time is a media cost with an accent.

What has shifted, then, is the question. The useful one is no longer