Workflow AI @WorkflowAIcom

Most companies are wasting money on AI. Not because the technology doesn’t work. Because they’re putting AI on top of broken workflows. They automate: → Tasks nobody should be doing → Reports nobody reads → Meetings that shouldn’t exist → Processes built around outdated assumptions Then they wonder why productivity barely changes. Before asking: “Where can we use AI?” Ask: “Why are we doing this work in the first place?” Eliminate → Simplify → Automate. Don’t automate a bad workflow. Redesign it.

@gregorojstersek So true — especially with AI. For example, if I use AI to turn meeting notes into an action plan in 10 minutes instead of spending an hour doing it manually, my manager only sees the 10 minutes. The 50 minutes saved is the real productivity gain — and I need to communicate that.

This one really hits home. As I’ve taken on more responsibility, I’ve realized that the goal isn’t to become better at doing everything myself. It’s getting to the point where someone says “I’ve got it” and you genuinely don’t feel the need to check. That trust takes time. You build it by being reliable, owning your mistakes, and doing what you say you’ll do. That’s when work starts to scale — when people trust each other enough to just get on with it.

I think the “intelligence becomes cheap” point is the really interesting part. If everyone can access extremely capable AI, then having access to intelligence stops being the competitive advantage. The advantage becomes knowing what to ask, what to build, which problems are worth solving, and how to orchestrate AI into a system that actually creates value. We're already seeing a small version of this today. Give the same AI tool to 100 people and you'll get wildly different outcomes depending on how they structure the work around it. AI may democratize intelligence, but it won't automatically democratize judgment, ambition, or the ability to turn intelligence into outcomes. So the real question might be: if AI makes intelligence abundant, what becomes scarce enough to determine who captures the value?

The fastest learner. Especially now that AI makes building and executing much easier. A great builder can build the wrong thing. A great seller can sell something nobody really needs. A consistent founder can consistently go in the wrong direction. The fastest learner can build → test → get feedback → adapt faster than everyone else. And over time, that compounds. The real advantage isn’t being right from the start. It’s learning faster than everyone else.

Most people don’t have an AI problem. They have a communication problem. They know what they want to say. They just don’t know how to say it clearly. Especially when the message is: → Asking your manager for a decision → Challenging an idea you disagree with → Escalating a problem → Saying “no” without sounding unhelpful → Communicating bad news → Following up when someone isn’t responding → Giving an executive update Here’s a simple AI workflow I use: 1. Write what you actually want to say. Don’t worry about grammar or tone. Write it exactly as it comes to your head. 2. Tell AI the context. Who are you writing to? What do you want them to do? What’s the sensitivity? 3. Give AI the outcome you want. For example: “Make this concise, assertive and professional. Don’t make it sound aggressive. Keep the message focused on the decision I need from them.” 4. Review the output yourself. AI should improve your communication—not replace your judgment. The biggest productivity gain isn’t getting AI to write for you. It’s getting AI to help you think about how to communicate. I’ve found this particularly useful for the emails that take 20 minutes to write but only need 5 minutes of actual thinking. What type of work email takes you the longest to write?

This is painfully accurate 😂 I’ve seen it a lot when using AI to improve workplace emails. You give it a simple message like “We won’t be able to meet the deadline because the material is delayed” and suddenly you get something like: “Given the current circumstances, we are navigating some challenges that may impact our ability to deliver within the originally anticipated timeframe.” It sounds polished, but it actually says less. I’ve started prompting AI to do the opposite: “Use plain, direct language. Say exactly what is happening, why it matters, and what needs to happen next.” The best AI writing isn’t the writing that sounds impressive. It’s the writing that makes the reader understand the point immediately.

I don’t think AI killed creativity. I think it made lazy creativity much easier. I’ve definitely caught myself doing this — asking AI for 10 ideas instead of sitting with a blank page and struggling to come up with one myself. But that struggle is often where the interesting ideas come from. AI can give you more possibilities, but having something worth saying still has to come from you. The danger isn’t AI making us less creative. It’s us becoming so comfortable with AI generating everything that we stop practicing creativity ourselves.

Honestly, I feel this too. When I started building a small digital product, I noticed I was using AI for almost everything — researching, writing, structuring ideas, even deciding what to do next. At some point I caught myself thinking: “Wait… am I actually building this, or am I just approving things AI gives me?” 😂 AI saved me a ton of time, but it also took away some of the fun of figuring things out myself. Now I’m trying to use it more selectively. If something is boring or repetitive, let AI do it. But if I’m learning, exploring or creating something I actually care about, I want to be involved in the process. Being 10x faster isn’t necessarily better if you stop enjoying the journey.

This is exactly the kind of problem that worries me with complex systems. A decision can look completely rational in isolation — slightly less safety for significantly more utility — while the combined effect of hundreds of those decisions may be very different. It’s similar to optimizing individual processes without looking at the end-to-end system: every local improvement makes sense, but the system-level outcome can still get worse. The hard part isn’t evaluating each trade-off. It’s understanding the interactions between them.

This is such a good leadership question. I started doing something similar with my team: when someone came to me with a decision, I’d ask “What would you decide if I wasn’t here?” Then we’d talk through the reasoning, define the guardrails, and I’d give them ownership of the decision. Over time, the questions changed from “What should I do?” to “Here’s what I recommend — does anything concern you?” That shift is huge. The goal isn’t to have fewer decisions. It’s to build a team that can make more of them without you.

The Mom Test. It changed how I think about building products. When I started experimenting with my own ideas, I realized how easy it is to get excited about the solution before proving that the problem is actually painful enough to solve. The biggest lesson for me: Don’t build what people say they want. Find problems they already care enough to solve. It’s saved me from spending a lot of time building things that sounded good in my head.

The open loops part is so true. A 2-minute task can take up way more mental space when you keep telling yourself you’ll deal with it later. I think AI can help here too — not by making every task faster, but by removing the friction that makes us postpone it in the first place. Close the loop. Clear your head. Give yourself some mental space back.

The “learning without practicing” one hits home. It’s easy to feel like you’re making progress because you’re researching, planning, or preparing. But at some point you have to put something into the real world and see what happens. Real feedback beats another hour of preparation. You don’t learn by preparing to start. You learn by starting.

This works really well for me too. When a task feels overwhelming, I can spend more time worrying about it than actually doing it. Giving yourself just 25 minutes makes it feel much more manageable. And once you get going, you often find that it isn’t as difficult as you imagined. Sometimes you just need to make the first step small enough to feel easy.

If AI saves you 2 hours of work, the answer shouldn’t automatically be “Great, now I can fit 2 more hours of work in.” Maybe those 2 hours should go toward a long lunch, a walk, your family, or simply doing nothing. Productivity should create more freedom, not just a higher workload. Optimize the work. Don’t optimize the life out of it.

“Close the loop” is probably the most underrated productivity principle here. Planning without feedback becomes a static plan. The real productivity cycle is: Plan → Execute → Measure → Learn → Adjust → Repeat AI can accelerate almost every step, but the feedback loop is what keeps you moving in the right direction.

The “ONE big thing” is the part I’d keep. AI makes it incredibly easy to create longer to-do lists, but more tasks don’t equal more productivity. The real leverage is deciding what deserves attention in the first place — then using AI to reduce the work around it. Prioritization first. Automation second.

A) Judgment. AI can increasingly generate ideas, communicate, and even simulate empathy. But knowing what matters, what to ignore, and what decision to make when the information is incomplete is much harder to automate. As AI makes execution cheaper, good judgment becomes more valuable.

3 small products. Not because 3 is better than 1, but because learning beats guessing. Three small launches give you three chances to discover: → What people actually want → What they're willing to pay for → What deserves to become a bigger product Start small. Let demand tell you where to go.