The classic amusement park “auto ride” lets riders drive a car — or at least “think” they are driving a car. The controlled car, mounted on rails, travels along a trough-like track and is powered by electrical currents. The driver can’t stop, accelerate or change direction — they can steer just enough to feel like they have some control over the car’s direction. Many companies also operate under a similar “command and control” model: A few company leaders hold all the power, specific company details (financials and salaries, for example) are closely guarded, and workers are expected to stay “on track.” Success is about following the rules. Compliance is valued more than trust.

While that controlling management style may work in an industrial context, it doesn’t work well in a knowledge economy powered by code. Modern organizations must go “off the rails” and empower workers with the data they need to determine their direction, speed, and destination. This empowerment mindset is crucial, as building ideas is very different from building physical goods.

Today’s business landscape is in a constant state of flux. The average tenure of a company on the S&P 500 was 61 years back in 1958. By 2016, that number had dropped to 24 years. At the current rate, about half of the list will be replaced in the next decade. Modern innovative companies are software-first, and software-first companies must embrace employee empowerment.

I’ve co-founded and led three previous companies, and each of those experiences taught me how vital it is to get culture right. That’s why I vowed to build my current startup on a foundation of complete trust and empowerment. I trust my tribe to make the right decisions at the edges of our organization. We have a strong focus on transparency, velocity, access to data, access to each other, and experimentation.

Your culture is your OS

A company’s culture is its operating system, and the OS is created and supported by everyone in the company. That cultural OS can even inform or define a company’s brand reputation. Adobe and Warby Parker’s respective work cultures have been called creative, collaborative, and innovative. Culture is the foundational bedrock that influences and defines a company from the inside out. Each day the company grows and adds new employees is a day that the culture is defined (and continuously redefined).

Aaron Dignan’s book Brave New Work (published in February) presents a model for the “brave new company,” and it’s a framework I urge company leaders to take a close look at. The book presents a 12-step program for exploring your company’s “OS canvas,” helping you articulate company purpose, structure, and innovation. As Dignan points out, all companies struggle with bottlenecks in decision-making, meeting overload, and siloed teams. But brave companies give everyone the freedom to choose when/where/how they work and make nearly all company information transparent. The brave new company makes it safe to experiment and gives everyone a voice in shaping the organization.

People want autonomy and authority; they want to understand the mission of their company and be empowered. This, in turn, drives innovation. Take Shopify, for example. The company has an internal “praise tool” that allows employees to offer cash bonuses to deserving colleagues. HubSpot is another company that espouses autonomy and freedom. Team members are encouraged to spearhead new projects while leaders at all levels work to be transparent, hiding nothing from teams.

How to go ‘off the rails’

So how do you create a people-forward culture in which people are trusted to do the right thing? What does this look like in terms of day-to-day company operations? How defined should company processes be? If “process is just writing down what works,” should we write down what works and then iterate from that? If we do, then the more defined a process is, the less flexible it and the company become. And as a company scales, this defined process can translate into more red tape.

Before creating your company’s “rail-free” culture, you need to make an important mental shift: Understand the difference between a complicated system and a complex one. A complicated system is similar to an engine or a watch – it’s hard to figure out, but with enough effort you can solve for it. A complex system, meanwhile, is like traffic or the weather: You cannot control all the variables because it’s a living, breathing, dynamic system. Today’s software-first, knowledge-based companies are complex, so it’s best to avoid endless rules and mandates in favor of trusting employees to make the right decisions. I think of it like a door. Sometimes you walk through a door and when you return to that door, it’s locked and you can’t get back into the room. Most people treat work-related decisions like a locked door; once a decision is made, there is no going back. The reality is that most decisions are like unlocked doors. We can walk back through the unlocked door. If a decision or action was wrong, we can learn, adjust, and try again.

Even though you’ll want to avoid rails and tracks at your company, you’ll need to design the fundamentals of your OS. Here are a couple of OS-level policies we’ve set up at my startup, for example:

A company-wide transparency default. Full transparency is our company’s default. That means everything — from partnership agreements to stock option grants — is made available by default to all team members unless someone consciously decides to go to bat to justify making specific information private.

Freedom to decide and act. We ask our team members to act first and ask for forgiveness later. No one should wait for permission to start a project or design a new product feature.

Financial transparency. We have full internal compensation transparency and full internal financial transparency. There are no wage surprises, as everyone who is a “software engineer 2” earns the same amount. Disparities are a non-issue, and this is especially important as it helps us ensure gender and diversity equality in pay.

Spending guidelines. We have a one-sentence spending policy: “When you spend the company’s money, do it in a way that’s best for the company.” Also, we have an “expense notices” channel, so everyone can see how we’re spending our collective funds. For example, I recently shared (via ExpenseBot): “I just saved $20 by sharing a Lyft ride from JFK to downtown New York and walking one block, about half the cost of a cab.”

As you can see, these OS guidelines are there to help build and support a collective conscience on our team, not to dictate detailed rules. They’re the guideposts or directional markers on the road, not the limiting tracks or rails.

Our company is always looking for ways to remove the rails and allow people to drive toward their destination on their own. Success requires learning, and learning requires failure. Our teams are free to try, fail, learn, and try again until they succeed. Our prime objective is to build a strong company, attract/retain talent, and enable the world to unlock innovation, through code. We have a culture of experimentation, and we experiment with our culture. We’ve scaled this approach through a Series B and our first 50 employees. Will it scale to 500? 5,000? I’ll report back in a year with an update, results, and lessons learned along our journey on the open road.

Daniel R. Odio is CEO of Armory.