The funnel is collapsing and agents are now buyers. Treat AI as a facilitator, not a finished product. And keep the human at the start of the creative.

It takes me back to my university days. I was a bioscience major at Carnegie Mellon, failing royally by year two, until an exam matched my personality to a career: science was at the very bottom, business, marketing, PR and communications at the top. As I leaned in, I became fascinated by the power of storytelling, through brand, through creative, and through data and measurement. I interned at CNBC at NBC Universal in ad sales marketing, then started my career at Bravo in the same division. I wasn't searching for B2B marketing, but when it came into my life I loved marrying the creative side of my brain with the revenue and business side.

Maybe it's full circle, but I'd call it a new evolution. With every emerging technology, storytelling becomes a new art, which is why I always say a marketer has to be a student again, and I am very much an active student. You have to learn the technology and consumer behaviour together, and being able to maximise the two and deliver a brand story with sharper edges is a fascinating challenge, especially as AI emerges around us.

I've always found it fascinating when new technologies move not just consumer behaviour but whole industries. I started in linear TV at the cusp of the dotcom age. At SyFy we gave away digital banners for free as added value, because there was no way to value them yet. Then came mobile advertising and my time at Apple defining in-app advertising, the millions of data points across TV and the app ecosystem, then CTV and streaming, where I spent 11 years defining the category.

To stay abreast, I look at what's emerging around us. I must have had 50 to 60 different conversations across founders, VCs and PEs today. That's my education right there. I love promoting new founders and businesses, and I always think about who I want to partner with next.

I use three criteria. First, privacy, data, governance and ethics as one, because partnerships fail when two parties hold a different ethos, and the integrity of the data and the tech has to be there. Second, is it really going to give us an edge and make a material impact, or is it just another shiny object wasting dollars, time and learning? Third, I want to be first all the time, first in the category, first to introduce a product, first to be a new measurement partner.

There's sometimes a blanket trust. You see the demo, the input and the output, but you don't look under the hood. At an enterprise level you have to look under the hood and ask the hard questions, because once you execute and partner together, you are as responsible as the tech that underlies it. The other mistake is deploying too quickly. We relied on the foundation of the business, built on the data we had, and used data-backed thought leadership to build sustainable long-term trust.

This morning I hosted a breakfast with Women Applying AI, a nonprofit I sit on the New York leadership hub for, 2,700 members across 10 cities, building AI fluency at any stage of your skills. In the founder's world I believe it's still male-dominated, and the stats will show that. There's still a sentiment that the tech space is meant for the tech bros.

I have an unproven and probably contentious thought: I think AI might level the field in terms of gender representation, because now you can have a one-person company and run a profitable business. I've met founders who are a team of one or two, with lots of smart agents and a really good idea. It has to start with a good idea, and a problem that needs fixing. I'm thrilled to see what that brings for women, men and different generations.

More than AI eroding tech and product development, I wonder whether we're looking at AI through the wrong lens. We assume AI companies have a product, but what if AI is a facilitator, helping new businesses emerge, accelerate, or existing businesses pivot? Someone told me the AI we're working with right now is the dumbest AI we'll ever work with, and it smacked me in the head. Anyone who thinks AI is a product will have to start thinking about AI as a facilitator. You'll see existing technology that underperformed come through via AI in less than six months.

I've debated this at least a dozen times in three months. The funnel is collapsing, awareness is collapsing. I asked a room of tech leaders once: does brand with a big B still matter? Of course brand matters. But put it in a matrix of commodity, price points and 1,800 competitors, and do you just buy the first three that are recommended? You now need a multi-tentacle marketing strategy. You still convert, reach and retain customers, but now you also have to think about the agents, which may be the more direct way of converting a client.

Reviews matter more than ever, and communities like Reddit play a powerful role. Until there's clearer governance over citations and fraud, the louder you speak, the more you get served. Speak to your core customer, don't betray your brand, and showcase the problem you solve. Think beyond the brand umbrella to everywhere the content could live, splice and customise the creative, and serve it in the right moment, whether through GEO or a creator. In my world, with mostly B2B, I always say market to demand first.

Websites have changed already, and I'll say they take on a new purpose. A website houses content, an arsenal of material you put out through all your marketing, that you can bring a client, partner or sponsor back into. Content can live anywhere. You still need a house, a marketing house that holds who you are, especially for press and reporters. It serves a purpose, but I'm not sure it's going to be required. Conversion can happen in the cart or through an Insta story.

My friend Anda Gansca, CEO and founder of Knotch, is on the cutting edge, optimising the customer path. A customer needs to see something five to six times before they even think about engaging, and for sales it's 10 to a dozen times. So you optimise not just the message but spot where a customer falls off the path, and imagine a dynamically optimised website that fits the journey you mapped. Something I've gone back to since 2015 at Samsung Ads is making sure your brand shows up organically through the customer's path.

The headlines say creatives are the first to be cut, that you can make a CTV ad in an instant. Those are headlines. I believe AI will help the creative evolve, and the creatives who last are the ones who use the technology, but the brain power, the thinking, the design and the tonality still have to be human. When you receive an AI content ad today, there's a feeling, no matter how realistic, something is still missing. It lacks a story. The human has to stay in the mix, and has to be the start.

Right now there's no clear governance around AI, period, and certainly not in agentic advertising. With every emerging tech, governance always came after. But this is a unique case where the tech is accelerating faster than we can handle, so my concern is that governance and regulation will be too far behind. There's a whole question of protecting kids, students and adults from fraudulent information, and we're being served content that abides by no governance, because there are no laws.

I was told recently that humans have a deeper trust with machines, because it's not another human judging them, no face giving a look. I find that disappointing and sad for the human race, but I can see the upside, in therapy or an honest dialogue you're uncomfortable having with a person. It was disheartening and exciting at the same time to hear we might be more trusting with a machine. My advice to a young professional: you have all the tools at your fingertips, so lean in, be a student, and don't be afraid to try.