The Future of Sales: Managing Virtual Agents Efficiently

Jason M. Lemkin discusses how AI sales agents will scale to 100-300 per user at $10-20 monthly, forcing teams to manage dozens of inboxes instead of a handful. The catch: enterprises face vendor sprawl, security audits, and renewal chaos, while the question of who actually pays for these agents remains unsolved.
Chapters
they'll probably buy like 100 agents and those agents will do the job of the human. They're going to pay for each agent $10 or $20 a month and they're going to manage them individually. The IT team or whoever is going to have a whole nother set of 200 or 300 agents they have to manage.
Yeah, I think— well, I'm not saying you're wrong.
I'm just trying to learn, actually.
Yeah, yeah, no, I think that's right. I mean, you think about like BDR of the future, right? BDR today manages one Gmail inbox with max 40 outbound emails. Maybe they manage 2 or 3. BDR of the future will be managing 40 or 50. Gmail inboxes. And so, um, you know, I think like if you think about like there's a deflation in the number of seats because now all of a sudden BDR acceleration, well, I don't need nearly as many BDRs, but the value per seat goes up in a pretty meaningful way. And so I think maybe what we see is application count stays flat or maybe even decreases, but the total amount of software spend increases because it's capturing some fraction of labor spend.
But are— so that, that's a— I, I get that and I can't predict the future, but aren't— if you just think about it from an old-fashioned enterprise, gotta manage X vendors, gotta manage rogue renewals, rogue increases. Like, do I really want an explosion of 5 times as many apps? Whether you call them agents, they're still apps from a, from a procurement IT process, aren't they? Yeah. Still another goddamn thing I gotta evaluate, do a security audit, do a cost-benefit analysis, make sure it doesn't go rogue. Like everyone went rogue in 2021 and bought 50 sales apps. Apps, right? I do— is— I just feel that the cognitive load is high. The co— not for the end user, but for it—
for the buyer. Yeah. I, I think the successful software applications, like, whatever, as a salesperson, I might have 500 agents, each of which is managing one lead. Yeah. You could. Right? And, and so what is that agent doing? It's gathering the news and the social media and the podcast and transcribing it. And that's pretty easy to review. I just go through my list, like, what's the latest with Databricks? What's the latest with Snowflake? And each one of those is an agent. So I don't think the cognitive overhead is huge if the UX is correct.
But who pays for it? That— I don't think for the— I think the bionic— I call it the mech AE, but the bionic SDR, I believe, is happening fast, right? But who's— I've never found a single AE on all of planet Earth or even our galaxy who will put anything on their own credit card. I've never found an AE.
More from Jason

Why Your Next Hire Should Report to an AI, Tragedy Apps, and Too Many AI SDRs? | The Agents #004

Our Agents Drove Our Salesforce Bill Up 80%, Marketo Hates Agents, and We Stealth-Churned Off Notion

Managing 20+ AI Agents: Lazy Agents, Our $500K AI Bill, Stealth Churn & the Death of 60% Solutions







