Former Google product leader: "I hired 40 AI agents for $500 a month, which is as effective as a $50,000 team."

A year ago, Jacob Bank had no marketing experience. Now, he operates 40 AI agents at a monthly cost of only $500, capable of doing the work of a marketing team worth $50,000.

"We should all have dozens or even hundreds of AI agents working for us. A 100-fold efficiency gap is not the future; it's the present."

The person who said this is Jacob Bank, a former Google product leader and now the founder and CEO of the automation platform Relay.app.

He's not just talking about the future—he's living in it.

In the past year, Jacob built a "virtual team" of 40 AI agents from scratch, which helps him with various tasks such as marketing, content creation, sales analysis, and competitive monitoring.

And he himself is the only marketer in the company.

I. Starting from Scratch: Don't Build 40 Agents at Once

"I was unlucky. I tried to build a single agent to handle 25 tasks and multiple jobs simultaneously. But later I found that wasn't the right way."

Jacob's advice is to start simple.

"Create an agent that can do only one thing. Create a second agent to do a second thing. Then, slowly, you might have an agent on top of them, calling them when necessary."

He gave an analogy: It's like hiring human employees. You wouldn't let a newly hired social media person be responsible for all social media marketing. You let them start simple.

"Don't start with the 40th agent; start with one."

Now, his 40 agents have their own responsibilities:

Automatically create LinkedIn posts every time a new YouTube video is published.

Automatically generate a tweet every time a new video is published.

Immediately notify him whenever a competitor's CEO posts on Twitter or LinkedIn.

Automatically check competitors' pricing pages every week and inform him of any important changes.

"So, I gradually built this AI agent team to help me."

II. "AI as an Intern"? That's the Wrong Mindset

Jacob admitted that he made a mistake a year ago.

"A year ago, I said, 'Hey, think of AI as your intern.' Then I realized that was the wrong mindset."

Why is it wrong?

"AI can certainly handle a certain type of tasks—those that save time and are usually given to interns. But AI can also conduct strategic analysis, competitive analysis, and content creation—things that interns can never do."

He gave his own example.

"I'm not good at cold calling and have no sales background; it's not in my nature. But it turns out that there are some very effective techniques in sales calls."

So, he created an AI sales coach.

This coach analyzes the transcripts of all his meetings and gives feedback: "You could have expressed this better. You jumped to the demo page too quickly. You should ask more exploratory questions."

"If you go out and hire a sales coach, a good one costs about $10,000 a month for one meeting a week. And you can build a pretty good AI sales coach yourself, which costs only about $5 a week including training."

AI is not an intern but a coach that can make you "super strong and super fast."

III. Future Work: Everyone Is a "Super Independent Contributor"

Jacob worked at Google for many years. There was a position called "Technical Lead Manager" there—a hybrid role that required both technical proficiency and team management skills. One had to be familiar with the entire codebase, review technical designs, solve the most difficult problems, and coordinate with others in the team.

"This is a very difficult role to play. It's hard to balance technical work and management work."

But he believes this is the future for all of us.

"We all need to become Super Independent Contributors (Super ICs)."

He defined future work distribution like this:

"Two-thirds of the day, I'm an individual contributor—I'm making YouTube videos, editing blog posts, publishing content, and talking to customers. One-third of the day, I'm coordinating work—my AI agent team helps me with that."

He gave an example to explain why low - end jobs are disappearing:

"If your job is a junior content marketer—rewriting a YouTube video into a blog post—that job will no longer exist. You can't build a job just by doing a single content repurposing task like this."

And high - end jobs are also changing:

"If your job is to manage a team of a thousand people in a large company, I don't want to say that job will disappear, but I think it will become less and less important. Because companies will be smaller. We'll see more lean, efficient, and smaller companies."

He made a bold prediction:

"Everyone will look like me—two - thirds of your day is spent on individual contributor work, and one - third is spent coordinating your AI agent team."

IV. AI Agents Are Not "Set - and - Forget"

Jacob repeatedly emphasized an important point: AI agents need continuous iteration.

"You can't just build an AI agent and say, 'This thing will definitely work for me forever.' I'm always modifying my AI agents."

He shared a real iteration process:

"After each evaluation call with a new customer, I spend 30 minutes talking to them about the application opportunities of AI in their business. I thought at that time, how cool it would be if they could get a personalized Google Doc with content I organized based on the specific cases and template links discussed in the call."

So he created an AI agent to do this.

"I said, 'AI agent, your task is to receive this transcript, organize the meeting content into a well - formatted Google Doc, and send it to the customer.'"

After doing it 10 times, he found the format was wrong, and the customers didn't even look at the document.

"I should have written that sentence directly in the email."

So he "fired" that agent and repurposed it for another thing: putting the summary directly in the follow - up email.

"Sometimes, you don't need a certain function anymore. Just like a year ago, we attached great importance to search engine optimization. Later, we realized that it wasn't an important channel for us—it only accounted for a small part of our new users and required relatively high effort. So we stopped focusing on search engine optimization."

"So I told my AI SEO agents, 'You can stop now. I'll contact you in the future.'"

What's the best part?

"There's no emotional burden. No disagreements. No coordination costs. I have an idea. Hey, AI assistant, can you help me implement this idea? They say, 'Yes, definitely.' Then let's do it."

V. "The Riskiest Career Is Stagnation"

Jacob has a subversive view of "risk."

"In my parents' generation, working at a large Fortune 500 company for 40 years was a stable career. You'd get a watch when you retired. And a risky career was either starting your own business or changing companies every few years."

"I think the situation has completely reversed."

"The riskiest career is when your skills are too closely tied to the environment of a particular company. This situation may change over time."

"And the most stable careers are often those of people who start their own companies. They have many different life experiences."

He took himself and his team as an example.

"Most of the members of our team are people I recruited from Google. Since we all worked at Google together, now we're starting this startup together."

"If you ask 10 people which is riskier: staying at Google or joining a startup? Most of them will say staying at Google is safer and joining a startup is riskier."

"As for your cash compensation next year, yes, Google will pay you more in cash next year. But what if you consider your career in the long run?"

"The key is how extensive a network of contacts you've built? How many new skills have you mastered? How do you push yourself to experience new things?"

"Staying at a company like Google is much riskier."

His conclusion is straightforward:

"I think the riskiest thing is stagnation. The risk lies in stagnation. Progress means learning, growing, and challenging yourself."

VI. How I Went from "Hating Social Media" to Being in Charge of Marketing

Jacob's experience itself validates his view.

"I have no background in marketing. I'm completely bad at it. Until a year ago, I hadn't even used any social media platforms. I hate sharing my private life in public."

"It made me very uncomfortable to be in charge of marketing."

But he realized one thing:

"Organic social media marketing, specifically on LinkedIn, is the best channel for our company's success."

So, he spent a whole year practicing hard.

"I spent a whole year practicing hard to learn how to do it well. Now I love it."

"So people might think, oh, this is too risky—putting so much effort into a marketing strategy we haven't tried before."

"No, this is an opportunity to leverage uncertainty—learning and growing in a new direction."

VII. If AI Agents Are the Future, What Should Our Children Do?

As a father of three children, Jacob often thinks about what kind of world the next generation will live in.

"I often think about how different the world my children will live in will be."

He lives in San Francisco, close to the test areas of Waymo and Zoox (Amazon's self - driving car division). Every day when he takes his daughter to school, she sits on the cargo bike behind him.

"They'll say, 'Waymo, Zoox.' For them, seeing an empty car is very normal. They think it's cool because there are a lot of cameras and lasers on it. But for them, the idea that a car can be driven by a person is strange."

"In the specific area of self - driving cars, I'm very happy. I think neither of them needs to drive alone. Because it's safer, more convenient, and better for the environment."

But in some aspects, his feelings are more complex.

"I don't want them to just keep swiping the screen. Even though YouTube's algorithm is very, very good, it still keeps pushing short videos."

He usually thinks from two aspects:

First, let the children experience real - world challenges.

"I try to let them experience as many real - world challenges and experiences as possible. You still have to learn how to ride a bike. You'll definitely scrape your knees when you ride a bike. I always tell my daughter, 'It's good for kids to have skin on their knees, right? Because if your knees are rubbing, it means you're trying something, taking risks in the real world, and dealing with the consequences.'"

"This was another major breakthrough when I was teaching my kids to swim—it's okay to swallow some water. This will definitely happen when learning to swim. So, I want them to take some risks in the real world."

Second, think about what skills will really be important in the future.

"I've spent a lot of time because I really want my children to learn logic and philosophy. Because I really hope they can express their thoughts very clearly."

He summarized two essential skills for the future:

"First, how well can you express what you want and tell it to AI? You need to clearly express what's important and what needs to be done, and then guide AI on how to do it."

"Second, how well can you build social relationships with others? Human abilities can be demonstrated in ways that AI can't. That's why I record our own YouTube videos—because it shows my personality, and people build social connections through it."

"I think these skills will be durable for any future career."

VIII. Facing AI, You Only Have Two Choices

Facing the current AI agent boom, Jacob believes there are only two ways to react:

First: "I'm skeptical. I don't believe it. I'm going to bury my head in the sand. I'll keep working the way I used to."

Second: "Wow, this is a great tool that will completely change the way we work. I want to be at the forefront of using this technology."

"Please choose the second one. It's more beneficial to you."

He emphasized: "I never want to tell you 'you should take more risks.' The way I present this discussion is: you should optimize your career to promote personal growth and learn new skills."

"Moreover, this is often the way to achieve personal growth and learn new skills—by entering a newer and more uncertain environment where you'll take on more responsibilities and be less certain about what will happen in the future."

"I don't think this is risky. I think the riskiest thing is stagnation."

"So for me, the key is to redefine the meaning of risk—the risk lies in stagnation, and progress means learning, growing, and challenging yourself."

This article is from the WeChat official account "Post - wave Evolution Planet", author: Mark. Republished by 36Kr with permission.