The Playbook Is Becoming Free

Over the past few years, I've sat through more agency pitches than I can count. Different logos. Different decks. Different names for proprietary systems. Eventually, I started noticing something. They were increasingly selling me the same thing.

One agency would identify a gap in our business. Another would come in a few weeks later and identify essentially the same gap. Both had a workflow to fix it. Both had an AI stack. Both had a diagram showing how much more efficient we could become. The terminology changed. The underlying ideas rarely did.

I don't think this means the agencies are necessarily bad. I think it means something much more interesting is happening.

The playbook is becoming free.

For most of business history, knowledge itself created an advantage. A consultant knew something you didn't. An agency had a process you hadn't seen. An expert had spent twenty years accumulating knowledge that would take you twenty years to recreate. That information asymmetry was valuable.

The internet started eroding it. AI is accelerating the process dramatically.

Today, two people can ask the same model how to improve an outbound sales process, structure a retention campaign, automate customer support, analyze a funnel, or build a content engine. They will get surprisingly similar answers. Then they can ask the model to create the workflow. Then the SOP. Then the emails. Then the prompts. Then the dashboard. Then the automation connecting all of it.

What once required years of accumulated institutional knowledge can increasingly be reconstructed in an afternoon. And because everyone is drawing from many of the same models, tools, articles, podcasts, posts, and case studies, something strange is happening:

Best practices are converging.

Everyone knows the funnel. Everyone knows the framework. Everyone knows the optimization. Everyone knows the workflow. And increasingly, everyone has the software required to execute it. Access to the playbook is becoming less valuable precisely because everyone has access to it.

I don't mean this pessimistically. The efficiency AI is creating is real. A small company can now perform work that would have required a much larger organization only a few years ago. Research that took days can take minutes. A first draft appears instantly. Thousands of customer conversations can be analyzed at once. Repetitive administrative work can disappear entirely. Software can be built faster. Experiments can be launched faster. A person can operate with capabilities that previously required an entire team.

This is extraordinary. But technological progress usually doesn't eliminate competitive advantage. It moves it.

When calculators became ubiquitous, being able to perform arithmetic quickly became less valuable. When spreadsheets became ubiquitous, financial modeling became dramatically easier.

But Excel didn't create a generation of Warren Buffetts.

It made calculation cheaper. Everyone could build the model. The advantage moved somewhere else: knowing which company was worth analyzing, which assumptions actually mattered, and when the numbers were telling you something the spreadsheet couldn't.

AI is doing something similar, except across a much larger portion of knowledge work.

If AI can write the code, deciding what to build matters more. If AI can create fifty advertisements, knowing which idea is worth advertising matters more. If AI can analyze every competitor, knowing which competitor actually matters becomes more important. If AI can produce a strategy, understanding whether that strategy makes sense for your particular business matters more. If AI can automate a workflow, knowing whether that workflow should exist becomes more important.

The scarce resource moves upstream. From execution toward judgment. From production toward selection. From knowing how toward knowing why.

This is why I think the operator becomes more important in an AI-native world, not less. The tools may become identical. The outcomes won't.

There is currently a Wild West around AI automation. Almost every company has inefficiencies, which means almost every company has something that can be automated. Find the manual process. Map the steps. Insert AI. Reduce the labor. Measure the savings. There is enormous value in doing this well.

But knowing how to automate something is very different from understanding the operation you're automating. You can automate a terrible process. You can make a bad customer experience happen faster. You can remove a human who was quietly handling exceptions nobody included in the workflow diagram. You can optimize a metric that shouldn't have been the goal in the first place. The automation can work perfectly while the business gets worse.

That is the part of AI implementation I think gets underestimated.

AI doesn't fix a bad operation. It lets a bad operation run faster.

The quality of the output is still constrained by the quality of the system surrounding it.

When someone shows me an AI workflow now, I'm becoming less interested in what it can do. Most of the time, I assume the technology can do it, or soon will. I'm more interested in why we're doing it.

Why does this process exist? What happens economically if we automate it? Where does it fail? What happens when the model is wrong? Which exceptions require a human? Does making this process faster actually improve the customer experience? Could we eliminate the process entirely instead?

These questions aren't as exciting as watching an agent complete twenty browser actions by itself. But they're usually more important. One of my favorite questions is becoming:

What did you decide not to automate?

That tells me far more about someone's understanding of an operation than a list of everything they managed to automate. Anyone with the right tools can identify things AI could do. Judgment is knowing what it should do.

Healthcare is one of the clearest examples. There will be AI for intake. AI for documentation. AI for patient communication. AI for scheduling. AI for clinical decision support. AI for personalization. AI for follow-up. Many of these systems will create enormous value.

But healthcare also exposes the limits of thinking about AI as a layer you simply insert into a workflow. Imagine a bad patient journey. The patient provides the same information three times. Nobody has complete context. The handoffs are confusing. Results arrive without explanation. Follow-up depends on the patient remembering what to do next.

Now automate all of it. You may have created an incredibly efficient version of a terrible experience.

The harder work is understanding how the experience should operate in the first place. Then technology can create leverage. This distinction matters far beyond healthcare.

Automation should follow understanding, not replace it.

This brings me back to all those agency pitches. I don't care nearly as much anymore if someone shows me a workflow I've seen before. Of course I've seen it before. Increasingly, everyone has. What I care about is the person operating it.

Do they understand the business? Can they recognize when the standard playbook doesn't apply? Can they distinguish a symptom from the actual problem? Can they make decisions when the data is incomplete? Can they recognize something technically correct that is practically stupid? Can they adapt when reality doesn't match the diagram? Can they decide when to ignore the playbook entirely?

Those abilities are difficult to package into a template. They come from judgment. And judgment comes from operating. Making decisions. Getting some wrong. Watching what happens. Correcting. Learning what matters. Doing it again.

AI can give someone access to an extraordinary amount of accumulated knowledge. It cannot give them the experience of having been responsible for the outcome. At least not yet.

I think we are heading toward a world where access to powerful AI stops being interesting. Everyone will have excellent models. Everyone will have agents. Everyone will have automation. Everyone will be able to generate software, content, analysis, research, and strategy at extraordinary speed.

That won't eliminate differences between companies. It may expose them. Because when two companies have access to increasingly similar tools, differences in outcomes become harder to blame on access. The difference moves elsewhere.

Which company asks better questions? Which one understands its customers better? Which one knows what to automate? Which one knows what not to automate? Which one moves when the answer is unclear? Which one has better taste? Which one can actually execute?

The tools become the baseline. The operator becomes the variable.

Every major technology eventually stops feeling like technology. Spreadsheets became spreadsheets. The internet became the internet. Cloud computing became infrastructure. AI will eventually become another assumed capability of doing business. When that happens, saying your company "uses AI" will sound about as meaningful as saying your company uses email.

The advantage won't come from having access. It will come from what you do with it.

That's why I don't think the commoditization of knowledge makes operators less important. I think it does the opposite. When information was scarce, knowing the playbook was an advantage. When intelligence becomes abundant, knowing when to use it, when to change it, and when to ignore it becomes the advantage.

Soon, everyone will have the tools. Everyone will have the workflows. Everyone will have the playbook.

Then we'll find out who can actually operate.

Subscribe to keep up.

See you Mondays, Maximilian