What's up, it's Zayd.

A few weeks ago I went to a meetup for founders and GTM leaders who are active on LinkedIn.

I usually resist going because they all sort of blend together, but this time I was curious.

At some point I mentioned that I don't use Claude Code for every single thing, and the room reacted like I'd offended someone.

The reaction surprised me.

We all have a whole army of tools that we use, so why was me saying that I am not using the star quarterback tool du jour for everything on par with calling the earth flat?

I feel like there's a version of the AI tools conversation that we have over and over: which tools are best, how to stack them, how to get the most out of each one.

That conversation is genuinely useful, but there's another version that almost nobody is having, which is when you should put the tool down entirely.

That's the one I want to have today.

Forwarded this email? Get Enabled in your inbox each week.

My favorite finds of the week

The pattern is predictable. A new hot AI tool hits the market.

Everyone’s a bandwagon fan (I include myself there 100%).

People rush to integrate it into everything they do. Output increases, speed increases, it feels like progress.

Then gradually, the quality of judgment in certain areas starts to erode because the tool has been making decisions that a person should be making.

Nobody notices until something goes wrong, and by then it's become a habit.

Using AI well is less about knowing how to use the tools and more about knowing when to use them.

Those are different skills, and the industry has spent almost all of its time on the first one.

The things AI is genuinely exceptional at: research synthesis, first-draft generation, processing large sets of information quickly, pattern recognition across data.

The things that suffer when you hand them to AI: nuanced judgment calls, relationship-specific decisions, anything where the context is inside your head and not on a page.

At the end of the day, no matter how hard you train your computer brains, AI can’t accomplish the hard things that require the kind of judgment you've spent years developing.

At Valley, AI handles the parts of a job where speed and scale create real leverage, and humans stay in control of the parts where judgment and relationship context are irreplaceable.

Our approval flow uses is a direct expression of that philosophy. Valley researches, Valley drafts, the rep reviews and clicks approve. The human just stops doing the work that was never the highest-value use of their time.

That structure matters because it keeps the rep's judgment in the loop without asking them to do the grinding work that AI can handle better.

The inverse failure mode is equally real. Teams that use AI to draft every customer message, generate every strategic decision, and shortcut every research process end up with output that is technically impressive, but fundamentally hollow.

A prospect can feel the difference between a message that came from a human who understood their situation and a message that was generated from a prompt. That gap is about whether there was genuine judgment behind them.

Examples of the best performing LinkedIn outreach messages:

The founders and sales leaders who will win in an AI-saturated market are the ones who develop a clear framework for which decisions belong to AI and which belong to a human being with context, relationships, and stakes in the outcome.

That framework looks different for every team, but it always starts from the same question: what is the irreducible human judgment here, and am I preserving it? If the answer is yes, use every AI tool available. If the answer is no, that's your signal to pull back.

The people who got the most out of the calculator were the ones who understood math well enough to know when the calculator was the right call and when they needed to think it through themselves.