De-risk the Company

Something can always kill the company

The principle, stated as one point in Alex Bouaziz's scaling playbook for Deel, is that there is always something that can kill a company and the founder's job is to de-risk it continuously. It is framed not as a one-time audit but as a discipline whose content changes with the company's stage. The governing distinction Bouaziz draws is between aggressive risk and fatal risk: "early on is the best moment to start taking risk. Never take risk that would put you or the company in jeopardy, that's a very bad idea."¹ The line runs between a bet that can hurt and a bet that can end the company, and the rule is to take the first freely early on while never taking the second.

Aggressive early, systematized late

Bouaziz gives concrete examples of risks he took early and would not repeat at scale, including promising a customer a missing feature "in three days" contingent on a deal and then building it in three days, and hiring people with zero experience to run functions, which "sometimes blew up in my face big time, sometimes it paid off."¹ The one domain where he says he never took shortcut risk, even early, is compliance and financial-services infrastructure: "I would always do it right, it costs a lot of money down the line, so we were always in the business of getting this right from the get-go."

As the company grew, the discipline systematized. After the collapse of Silicon Valley Bank, Bouaziz says Deel keeps "three backup banks" for every bank it uses, on the logic that a country can change regulations overnight or a bank provider can be acquired and drop you while payroll still has to run, a risk category he calls "just financial services."¹ The scaled apparatus he describes also includes eventually hiring a chief risk officer once the risk surface justifies a full-time owner, explicitly mitigating key-person risk where "a couple of people who know everything about X, Y, Z" would leave the company exposed if they departed, and a cadence in which "once a quarter you write down the 10 biggest risks, prioritize them, and decide which to tackle."

The epistemic limit built into the method

The most striking part of Bouaziz's framing is that he names the method's own gap. "Sometimes you don't even know things are risks until they're risks, that's the hardest part," he says, observing that no company's quarterly risk plan had listed the failure of Silicon Valley Bank.¹ The quarterly top-ten catches only known risks, while the events that actually kill companies are frequently unmodeled. His partial answer is structural rather than enumerative: redundancy such as three banks per bank protects against a class of shocks you did not specifically forecast, which is why he leans on backups rather than trusting a risk list to be complete.

How it relates and where it stops

The principle pairs directly with worst-case scenario first, the mindset meant to surface the unmodeled tail before it arrives, and with capital allocation discipline, since Bouaziz treats a low-burn, EBITDA-positive balance sheet as itself a de-risking move that removes the run-out-of-cash failure mode.¹ It also sits close to a wartime operating posture that treats constant threat as the baseline. The principle's honest limitation is that it cannot close the gap it identifies. Structural redundancy is a partial answer, but a company cannot triplicate everything, and the clean rule to take aggressive but never fatal risk offers no way to tell an aggressive bet from a fatal one in the moment, which is exactly the judgment the rule requires and cannot supply. The concept is therefore best read as a staged appetite for risk plus a concrete practice, not a guarantee against the shocks nobody wrote down.