I saw a LinkedIn post the other day from someone who creates content for B2B software startups, arguing that all salespeople should use voice agents to automate their cold calling. I’ve heard some version of this same argument dozens of times in the past few months, and I understand it on the surface. AI does a really good job at bringing efficiency to existing functions, and cold calling is a numbers game where more input leads causally to more output.
What bothers me about this line of thinking is that it treats AI as a productivity layer that is being applied to a static landscape. In reality, AI is changing the landscape so fundamentally that old functions shouldn’t be automated, but done away with. The same technology that makes a voice agent possible is also changing how buyers research, evaluate, and decide. They're doing more of the work themselves, earlier, and asynchronously. They expect the people reaching out to them to already know who they are and what they care about. Cold calling was only ever tolerated as part of the cost of doing business in a world with limited information signals, but we don't live in that world anymore.
Rather than using agents to make more calls (or send more emails, or publish more ads), the actual opportunity sitting in front of GTM teams lies in the explosion of signals. It’s easier than ever before to understand who's reading what, who's in market, who's been to your site three times this week, or who's talking about your brand in their Slack community. The value isn't in automating the input, but in learning to use the signals to be more precise about when you show up and what you say.
There's also a shift happening around permission. Outbound used to operate on an implicit permission: you were a prospect, so outreach was expected. That's no longer true. Permission now has to be earned by being useful before you ask for someone's time, by showing up in the places they already pay attention, or by being championed by people they already trust. There’s no silver bullet or shortcut for any of these processes, via a voice agent or any other sort of AI.
The moat doesn’t lie in optimizing or automating old motions, but in building new ones off the back of the opportunity that AI creates. Good marketers today are embedding themselves in the communities their buyers are already a part of and partnering with organizations and people who've already earned trust in those communities. They're producing things that their ICP actually wants to consume. AI certainly plays a role in developing these new motions, but it’s not a replacement for deeply understanding your customers.
Doing this right requires patience and conviction, which is the hardest part for most organizations. The ROI of owning your own community, or executing a large-scale brand activation, is lagging and hard to capture in traditional attribution models. The work of identifying which organizations have already built trust with your ICP, and figuring out how to partner with them, requires sustained attention and a fair amount of taste. It’s certainly much harder than buying a lead list or running a paid LinkedIn campaign. But, at the end of the day, it’s marketing’s job to tell the story about why this “slower” work is so important. Teams that make the case effectively will build real moats as their work compounds over time. Once you've embedded in a community and built relationships with the people who matter in it, you have something competitors can't replicate.
