The Easier It Gets to Build, the More Important It Is to Know What You’re Building
September 17, 2026
The cost and difficulty of building new capabilities are falling quickly.
Software can be prototyped faster. Data can be analyzed with less specialized effort. Workflows can be automated more easily. Ideas that once required months of development can often be tested in days or weeks.
AI is one reason for that change, but the broader implication matters more: when building gets easier, deciding what should be built becomes more consequential.
That is why I think systems thinking is becoming more valuable.
Start with the target state
For me, systems thinking starts with a simple question:
What should be different when this works?
That sounds obvious, but organizations often skip it. They move quickly from a problem or an opportunity into projects, tools, vendors, or implementation plans.
The better starting point is the target state.
What should the customer be able to do? What should employees be able to see or decide? What information needs to move from one part of the business to another? Which parts of the process should be automated? Where does human judgment matter? Which systems need to connect? What capabilities already exist that could be used differently? What new capability is genuinely required?
And ultimately: what business result is the whole system supposed to produce?
Once the target state is defined, you can work backward from it. You can identify the capabilities, dependencies, decisions, economics, systems, and behaviors required to make it real.
That is very different from starting with a technology and asking where it might fit.
Local improvement is getting easier
One of the most useful things new technology can do is improve a defined task.
Take something that requires two hours of manual work and reduce it to twenty minutes. That is real value.
But businesses are systems. Improving one step does not guarantee a better result.
You can generate more leads without improving sales. You can automate a handoff without fixing the fact that the person receiving it does not have the information needed to act. You can create more analysis without improving the decision. You can make a bad process dramatically more efficient without asking whether the process should exist in its current form.
The larger opportunity appears when you start with the desired business state and then ask what combination of people, process, data, and technology would make that state possible.
I have seen this pattern before
This way of thinking predates the current AI wave.
Years ago, I was working on a major retail expansion problem. The company had a large list of possible markets and an aggressive growth plan. Useful analytical capabilities already existed inside the business: a financial model for potential stores and a geographic sales-forecasting capability.
The important move was not inventing either one. It was recognizing that they could be combined and applied across the entire market universe.
That changed the question. Instead of evaluating locations one at a time, we could understand the size and shape of the expansion opportunity as a whole.
The technology enabled the analysis. The value came from seeing how the pieces could work together.
I saw a similar pattern in Army recruiting. The recruiting journey crossed marketing, a contact center, recruiters, appointments, applications, and enlistment. Different systems and organizations touched different parts of that journey.
Once we looked across the full system, new possibilities became visible. Existing data could help infer where someone was in the process. Information collected in one interaction could become useful to a recruiter downstream. Communications could change based on what had already happened.
Some of those ideas could be tested with capabilities that already existed. That made it possible to launch MVPs, learn from the results, and use evidence to support further investment.
In both cases, the important skill was not knowing how to engineer every component. It was recognizing what should be possible and defining the system well enough that technical teams could make it practical.
This is where AI changes the equation
AI increases the importance of that skill because it lowers the implementation barrier.
If a team can generate code faster, build a prototype quickly, automate a workflow, or interrogate data without waiting for a long development cycle, more ideas become feasible.
That is useful. It is also risky.
The cheaper it becomes to build something, the cheaper it becomes to build the wrong thing.
The leadership challenge shifts upstream. Someone still has to frame the problem, define the target state, understand the dependencies, and decide which capabilities should exist in the first place.
The World Economic Forum's Future of Jobs Report 2025 points in the same direction. Employers expect systems thinking to become increasingly important through 2030, and the report connects that rise to technological change and the growing interaction between people and automated systems.
Deloitte has made a similar argument in its work on AI-first organizations, identifying systems thinking and problem framing as important capabilities for work that increasingly spans people, tools, and automated systems.
There is also reason to be careful about outsourcing the framing step itself. An INSEAD working paper found that large language models helped people generate more alternatives, but that strategic focus dropped when the models were also used to frame the problem in the first place. When they assisted only with generating ideas, that effect did not appear. The implication is not that leaders should avoid AI. It is that the quality of the framing still shapes the quality of everything that follows.
Technical fluency still matters
None of this means leaders can ignore technology.
They need enough technical fluency to understand what a technology can do, where its limits are, what questions to ask, and when a proposed solution does not make sense.
But that is different from needing to know how to build every component themselves.
As the tools become more capable, the premium moves toward people who can connect those capabilities to the broader business system.
They need to be able to answer questions such as:
What are we trying to make possible? What should the target state look like? What has to change in the process, organization, data, or systems? What can we test with what we already have? Which new capabilities are worth investing in?
Those are not engineering questions. They are leadership questions.
Better tools make better framing more important
For a long time, one of the practical limits on changing a business was the effort required to build the solution.
That limit is weakening.
As implementation gets easier, the quality of the thinking that comes before implementation matters more.
The organizations that benefit most from new technology will not simply be the ones that can build faster. They will be the ones that can define the target state well enough to know what deserves to be built.