A network’s value comes not just from access, but from trust, judgment, and signal, which is why automating introductions at scale can destroy what makes networking useful.
“It’s not what you know, but who you know,” according to the common refrain. But what does it mean to “know” someone? I can collect tens of thousands of email addresses or connect to thousands of strangers online, but is that knowing them? Are they really in my network? The true value comes from being able to rely on your network because of the trust that has been created. When someone introduces you to someone else in their network, they make a judgment call about their mutual value to both parties. That judgment isn’t something AI can easily replicate today.
In “Automating Networking with AI – Making Networking More Efficient,” I covered how AI could be used to automate staying top of mind in our networks. More generally, AI also automates the task of searching across our extended networks. But doing so raises the question: what exactly gives a network its value, and does that value persist without that human trust and judgment?
Consider the following common case. If you’re looking for a job, you could contact your friends and ask if they know anyone hiring for your job category. They will try to think of anyone they know who is hiring. Alternatively, they might say, “Look through my network and if you see anyone hiring for your role, let me know and I’ll connect you.” This is exactly the type of grunt work modern AI does well; it can parse through pages of unstructured text, like LinkedIn profiles and posts, and see who is hiring.
Let’s go one step further. In “Automating Networking with AI – Making Networking More Efficient,” I discussed how AI agents could automate staying top of mind. Imagine that you click a button that says “find a new job” and it reaches out to all the people in your network to let them know, and each of those people has an agent search on your behalf. (Note: I’m assuming their agents do the work since it scales exponentially when you get to second-degree connections, but it works the same either way.) When your contact’s agent finds someone in their network who is hiring, it sends an introductory email, or forwards your resume automatically. It’s networking on steroids!
But here’s the catch. I’ve long argued that you should only add people to your network whom you genuinely know. I’ll only consider adding someone to my LinkedIn contacts (a proxy for my true network) if I’ve had some level of interaction with them that was net positive (even if just mildly positive). Of course that doesn’t mean I’d recommend everyone in my network to everyone else in my network. I use human judgment about who I’ll introduce to whom. I have said no to vendors who simply wanted to connect to senior executives in my network when I know those executives do not want such inbound requests. That’s the judgment part of networking. It’s why when I do make an introduction, people know it carries weight.
Other people take a different approach; they’ll connect to anyone and have tens of thousands of connections. In a world where many people do this (perhaps by automating the outreach as described above), almost anyone becomes reachable and it can all be automated. When you need a job, for example, AI can look through this extended supernetwork and find everyone hiring for that role and make an introduction. If you automate this step, suddenly introductions start flying around. Instead of being a filter, your network becomes a source of spam.
But think of the implications. The hiring manager, instead of getting one or two highly recommended candidates through her network, now gets hundreds of “recommended” applicants through her supernetwork. The network has lost much of its value (other than perhaps discovery); there is no benefit to an introduction made in such an environment and no efficiency gain because it no longer provides a strong signal among the noise.
There have been a number of services over the years where you can pay someone who will introduce you to their connections. Not surprisingly, none of these services ever did very well, because the best high-value relationships are not for sale, and would lose their value if they were. This AI-powered networking-on-steroids approach leads to the same result.
In the infinite monkey theorem, it is suggested that enough monkeys sitting at keyboards over time will eventually replicate great works of literature. In theory, if enough monkeys (Homo sapiens instead of Pan troglodytes) and their AI agents reach out to enough people, it could unlock a huge amount of value currently held back by the friction of networking.
The infinite monkeys were only a thought experiment until 2002 when students from the University of Plymouth made it real by putting a keyboard into a monkey enclosure. The monkeys produced nothing of value and defecated on the machine (source). I suspect those who use AI to automate the relational and judgment-based parts of networking (as opposed to the mechanical side) will wind up with a similar outcome.
It’s critical to learn about corporate culture before you accept a job offer but it can be awkward to raise such questions. Learn what to ask and how to ask it to avoid landing yourself in a bad situation.
Investing just a few hours per year will help you focus and advance in your career.
Groups with a high barrier to entry and high trust are often the most valuable groups to join.
