I recently sat down with Sean Byrnes — founder, executive coach, and one of the sharpest GTM thinkers we know — to talk about what's actually working in go-to-market right now. And what isn't.

For those who missed the Livestream, here’s the full recording.

What followed was one of the most honest conversations we've had about the state of B2B sales and marketing in the AI era.

Here's what came out of it.

The AI buying window is closed

For a brief moment — 2024, early 2025 — buyers would purchase almost anything with AI in the name. They didn't have criteria. They didn't have expertise. They just needed to tell their boss they were doing something with AI.

That window is closed.

Buyers have been burned. Churn rates are high. The market is getting discerning in a way it wasn't eighteen months ago. The hype was real. The hangover is realer.

Everyone can build now — so nobody stands out

The cost of building software has dropped to almost nothing. That sounds like opportunity — and it is — but it creates a new problem. If anyone can build it, hundreds of people will. And they have.

Every category is crowded. Every buyer has already tried two or three versions of your product and has opinions about why they didn't work.

Sean calls these the “zombie hordes” – unviable or disappointing products everywhere, threatening your prospects.

Sean's framing stuck with me: "The zombie hordes aren't smart. There's just a lot of them. And millions of zombies is a huge problem."

This is the environment your GTM has to cut through.

GTM was always the hardest part

The ease of building has shown a new light on how hard GTM is. It’s not only that there are more options–it’s that you need to really differentiate and cut through to win attention.

It doesn’t help that in this AI-fueled building boom, not only are there more competitors, but also there is far more appetite to DIY.

You can’t dismiss DIY

Now that everyone can build, DIY has become a real competitor. Is your app better? Sure. But if someone at the company insists on rolling their own–that is still a competitive alternative you have to be ready to address.

Just because they don’t know how hard it is doesn’t mean they won’t end up in your “closed-lost” graveyard. Surrounded by zombies.

Positioning is key–but if you have a random mix of customers, beware

This one hit home for me: you ship a product, have lots of customers, but there’s no pattern. It’s hard to go deep on a specific vertical because, well, everyone is using it.

But it’s also hard to get deeper traction in any vertical because you are being pulled in multiple directions.

For Skyp this looks like mass-senders who don’t care as much about enrichment as they do about volume, sending 100k+ emails a month. These companies spend a small fortune on email–every month. But doing fully AI-enabled outbound would likely bankrupt them; it’s just too costly to go deep enough on each lead to make it pay off at this scale at mass market, lower ACVs.

Our original target, founders and early stage sales teams, don’t go for volume. They care about enrichment, targeting, and personalization. They want to invest in each lead, to maximize their chances of getting a response–because their initial market may have only 1,000 or 5,000 companies. At 100k emails a month they’d blow through it in a day.

Those are very different products–and we have both types of customers. Choosing which to serve well and which to serve less well has been a challenge. Not to mention that our customer base is spread across almost a dozen industries.

The step that many startups skip

Sean’s point is that you get to this mess of customers with no pattern because you skipped an important step: customer research. And there are good reasons why many companies skip it–but that doesn’t make it ok.

Companies skip customer research because building with AI is fast and fun. It used to take 2-3 months to ship a new feature. This came with real costs–in headcount, burn, time, and other resources. There were real tradeoffs–if you built this feature this quarter, that meant you obviously couldn’t build some other feature. So you had to choose.

To make these tough choices, the best companies invested in customer research. Sure it took a month or two, but given the investment required, it was obviously worth it.

AI changes that math completely.

Shipping a feature might take a day or two. A week would be a long time. Marcus Segal at a16z Speedrun talked about how they’ve shortened their “short term” product planning from 2 weeks to 1 week because of AI.

Today, the customer research takes way longer than just shipping an 80% baked product and seeing what happens. Plus we’ve all learned–shipping is more fun! Which would you rather do?

Because of that startups throw products and features at the market. Sean sees the same result across a lot of companies – what I said above, a random mix of customers with no way to go deep on any one vertical.

Founders then wonder why they can't build a repeatable GTM strategy and are trapped in a spiral. To get out of it, they hire a VP of Marketing.

Doesn't work.

Then a VP of Sales.

Doesn't work. The problem isn't the people. It's that nobody went back to do the work that should have happened before the first hire.

How to niche down effectively

People call focusing on one vertical “niching down”. Here’s how to get there.

Talk to every existing customer — not about what they use, but about why they bought. Specifically: what did they think this purchase was going to do for their career? Sean's thesis is that most software gets bought because someone wants to get promoted. If you understand that motivation — the person's, not the company's — you understand your real ICP.

Then reach the ones who bought from someone else. Not to pitch — to understand their buying process. Sean found that framing this as market intelligence opened doors that a sales pitch never would.

It's fifty to a hundred conversations. It takes months. Almost nobody does it.

It’s important because now anyone can ship a product. GTM is the hard part, and being successful requires getting GTM right more than it requires having the perfect product.

To get GTM right, you have to understand what that initial niche values and needs, how they talk about it, and how they buy. Only then you can build the perfect product for your target market.

What's working in GTM today: multimodal GTM

The companies growing right now aren't doing one thing. They're doing several — and the combination is what's working.

Outbound email. Conferences. Content. Events. Phone calls. The channels reinforce each other in ways that none of them can do alone. If someone hadn't gotten the email, they wouldn't stop by the booth. But the email alone wouldn't have been enough either. I’ve said this before about conferences, and it’s even truer today. .

Conferences have come back in a particular way that suits the AI era. AI products have something SaaS never had — a genuine "wow" moment when someone sees them for the first time. Fifteen seconds of real attention at a booth can drive a conversation that an email never would.

AI buyers are buying outcomes, not software

Buyers are asking: what's the headcount equivalent of this outcome? How many people would it take to do what this does? And they're willing to pay accordingly — often two or three times what they'd pay for comparable SaaS — if the outcome is real.

This changes your responsibility. You're not selling access to software. You're selling an outcome you have to actually deliver.

It also changes what buyers want from you. They're not looking for a vendor. They're looking for someone who helps them navigate a world that's changing faster than they can keep up with. Sean's framing: "They're not looking for you to sell them the answer key. They want you to help them with their homework."

The new rules haven't been written yet

The old playbook is broken and most people are still running it.

The new one is taking shape: do the customer research, build multimodal GTM, position around outcomes, and be a partner — not just a vendor.

The rules are being thrown out. That's the anxiety. But it's also the opportunity. The founders who figure out what actually works in this environment — not what worked in 2019 — are the ones who'll have built something defensible when the dust settles.

Next up

We have 3 more of these livestreams (don’t call it a webinar!) already lined up in April–I’m really excited about all of them. Hope you can join. To get notified, subscribe to this newsletter (I’m sure you saw a popup by now, if not here’s the homepage for it) or sign up on the Luma for any in the series; we send invites a few days in advance.

And if you want AI email to be part of your multimodal strategy, you know where to find us. Happy building + selling!