In a World Where Everyone Has AI, Systems Thinkers Win

🔥 The edge isn't the modelLink to this section

Every company can rent the same models. What sets some apart is a way of thinking.

Systems thinking comes down to five habits:

• See the whole problem, not just your piece.
• Go back to first principles.
• Spot patterns.
• Build reusable blocks.
• Design feedback loops so the work improves itself.

None of this is new. It just matters more now. As Netflix CPTO Elizabeth Stone put it on Lenny's Podcast:

"We need more systems thinkers in a world with AI."

The more AI connects data and workflows across an organization, the more valuable it is to have people who can see across those boundaries.

I saw all five habits at work at Knoq, long before AI made them obvious.

🚪 Start with the problem, not the productLink to this section

At Knoq, I led product for a political canvassing app. Reps knocked on doors and recorded what people cared about and how they leaned. Then we pivoted to selling home products and services.

The obvious question was how to turn a canvassing app into a sales app. We asked a more basic one: what has to be true for person-to-person sales to work?

Framed that way, a lot carried over. Reps still had to understand the person in front of them, adapt the conversation, and work a territory. What changed was everything around that. A campaign is a finite push; a sales org is an ongoing system of leads, reps, customers, and CRM handoffs.

So we kept what worked and rebuilt around the new problem. Lesson one: don't confuse what you built with the problem you're solving.

📍 The data was bigger than its original purposeLink to this section

Our reps collected information you couldn't buy or scrape: what neighborhoods cared about, how they leaned, the condition of the house, the car in the driveway.

We'd been treating political data, sales data, and neighborhood observations as separate things. But a signal from one context could be useful in another. So we stopped asking only "What does this tell us about this campaign?" and started asking "What else could this tell us?"

With millions of interactions, no person could find what mattered. So we built an engine to find the patterns.

🎰 Patterns nobody would guessLink to this section

One example: people who had voted in favor of casinos in their area were more likely to sign up for energy choice plans.

That pattern changed who we talked to, where we spent marketing dollars, and what reps said at the door. Reps didn't need to understand the model; the system just gave them better information at the right moment. And every conversation's outcome fed back in:

data → pattern → action → outcome → data

No analyst would have spotted that relationship. But the machine couldn't decide what it meant for the business, either. The value came from combining people who understood the system with technology that could see across it.

That was with the machine learning of the time. Now imagine a systems thinker with today's AI.

⚡ Why this matters more nowLink to this section

The same habits apply:

• See the whole problem. Find where AI creates leverage across the system, not just one task.
• Go back to first principles. Ask whether a process is right before automating it.
• Spot patterns. Connect information that was never designed to connect.
• Build reusable blocks. Turn what works into capabilities you can combine.
• Design feedback loops. Let outcomes improve what comes next.

Point today's AI at a well-designed system and the ceiling moves.

Point it at a messy one and you get more noise, faster.

#SystemsThinking #AIStrategy #AIAdoption #ProductStrategy #Leadership #FirstPrinciples #FeedbackLoops