Most Organizations Have the Wrong Strengths for the AI Era
Velocity and execution are rare, even, or especially, in technology companies.
Architected Intelligence: Principles for Building AI-First Organizations and Technologies by Jacob Miller and Jeremy Mumford is available now.
What makes a business succeed? Many frameworks reduce it to some variation of three things:
Market – You are competing somewhere worth winning. It is better to be a mediocre company in a great market than a brilliant company in a dying one.
Strategy – You differentiate in the right way and get the product into customers’ hands before someone else does. You give customers a reason to choose you and a reason to stay.
Execution – You repeatedly deliver higher quality at lower cost. You improve faster than the competition. You turn learning into a system.
Different eras reward different strengths. Buffett built Berkshire largely around market selection and capital allocation. Salesforce built an empire around strategy and workflow entrenchment. Toyota turned execution into a quasi religion.
For the last twenty years, execution has often been the quiet third variable. Many companies succeeded with strong markets and strong strategy while execution ranged from excellent to surprisingly mediocre. High margins, venture capital, network effects, and digital distribution covered a lot of operational inefficiency. The lesson many organizations internalized, even if they never said it explicitly, was that execution mattered, but it did not matter as much as being in the right market at the right time with the right positioning.
And that’s when AI showed up…
AI reduces the cost of doing work. That part is obvious and widely discussed. What is less discussed is that AI also reduces the cost of learning. Teams can build faster, but they can also gather feedback faster, synthesize it faster, analyze failures faster, generate alternatives faster, and iterate faster. The build loop accelerates and the learning loop accelerates. Both sides of the equation improve at the same time.
Because of that, execution should be thought of less as effort or culture and more as a measurable capability:
Execution is the number of high-quality build-and-learn cycles an organization can complete in a given period of time.
Once you think about it this way, the impact of AI becomes clearer. AI increases the number of cycles you can run AND improves what you learn from each cycle. The organizations that benefit most will not necessarily be the ones with the best strategy documents or the most ambitious AI roadmaps. They will be the organizations that can run more high-quality cycles than everyone else.
Many in the venture capital industry have been beating the drum making similar points. Marc Andreessen beats the drum on OODA loops (observe, orient, decide, act). At a recent investor panel, I heard Ethan Choi of Khosla Ventures repeatedly emphasizing that the primary attributes they look for in companies were the founding team’s ability to execute with velocity.
This is where the idea of the gingerbread man moat comes in. The gingerbread man does not win by being bigger or stronger or cuter. Instead, he wins by being difficult to catch. He keeps moving, and when he stops moving, he gets eaten. For many industries right now, that is the situation. Traditional moats still exist. Regulation still matters. Brand still matters. Distribution still matters. Capital intensity still matters. Data still matters. But velocity (execution) is becoming a moat available to far more companies than before.
The companies that can iterate, learn, and improve faster will feel very difficult to compete against because they keep widening the gap. Most organizations are not built for that.
The Recipe for Velocity
When you look closely at teams that execute quickly, a pattern shows up repeatedly. They tend to have more decision makers (without complete decentralization) with real authority operating closer to the work. That sounds simple, but it requires three things to be true at the same time.
Empowered – They are allowed to make decisions and they have the resources to act on those decisions. Authority without resources results in flailing. Resources without authority create frustration. Both authority and resources are required.
Aligned – They are accountable for the outcome, they believe in the direction, and their incentives point in the same direction as the organization. They receive feedback not only on the product, but also on their own decisions and performance so they can improve along with the system they are building.
Informed – They are close to the user, the product, and the technology. They understand the problems and the edge cases. They are NOT operating entirely through dashboards and summaries. They have proximity to reality, and they still bring a magnifying glass.
Empowered, Aligned, Informed. Each is necessary, and none is sufficient alone.
There is also a surprisingly simple way to get a rough sense of whether an organization is structured this way. Look at the meetings. Do not count the number of meetings. Instead, count the number of attendees in the meetings.
Meeting count is a noisy metric. A team can have many short meetings and still move quickly. Or, a team can have collaborative four hour working sessions with three people. The more revealing signal is how many people need to be in the room for a quality decision to happen. When empowerment, alignment, and information are concentrated in the right people, meetings tend to be small. When those things are fragmented across layers and functions, committees appear, either formally or informally. Committees meet because they have to. If decisions require pieces that live in different places (empowerment, alignment, or information), those pieces have to be assembled regularly. Often, they fail in their attempt to assemble each piece. The more fragmented, the more people are necessary and the likelihood increases that a critical component is missed.
Large recurring meetings are often a structural signal that suggests that components are distributed across the organization in a way that slows decision making and iteration. And now with AI, iteration speed is increasingly the game.
To compress this argument to three points:
AI has tilted the environment back toward execution. Not exclusively toward execution, but enough that execution capability is becoming more important relative to market selection and high-level strategy than it has been in a long time.
Many organizations optimized for the previous environment where market position, strategy, and access to capital carried more weight. They built structures that were good at planning, coordinating, and managing risk across large organizations. Those are still valuable capabilities, but they do not produce high iteration velocity.
The organizations that will do well in this environment are the ones that can empower the right people, align them to outcomes, keep them informed and close to reality, and then allow them to run many high-quality build-and-learn cycles.
Many organizations do not have those strengths and are not maniacally pursuing them. Over the next 24 months, that gap will weigh more than the pile of AI-written strategy documents attempting to dig the company out of its hole.
Architected Intelligence: Principles for Building AI-First Organizations and Technologies by Jacob Miller and Jeremy Mumford is available now. Published by Wiley, 2026.
