Here's a thought experiment for you: What if Elon Musk ran an insurance company?
Just imagine how regulators would react to his brash, visionary ideas wrapped in disdain for tradition and a belief that rules don't apply to him.
But what if you could bottle the good parts of his iconoclasm and leave out the parts that would scare policyholders about the reliability of their insurance carriers? A former president of Tesla just published a book that might allow for that. It describes the five-part "algorithm" that he and Musk used to manage the company during a transformative stretch in the mid-2010s.
I don't think insurers should go full force, a la Musk's "hardcore" mode--you could wind up with an embarrassment like DOGE and never recover--but his algorithm does offer a playbook for radical simplicity and for what I think is the right way to approach automation.
Jon McNeill, author of "The Algorithm: The Hypergrowth Formula That Transformed Tesla, Lululemon, General Motors, and SpaceX," says the method has five steps:
- Question every requirement.
- Delete every possible step in the process.
- Simplify and optimize.
- Accelerate cycle time.
- Automate.
Question Every Requirement
McNeill writes about how Tesla, for instance, questioned China's requirement that it own a piece of any company operating in the country and eventually negotiated a deal that let Tesla own 100% of its Chinese subsidiary. He also writes about deciding that cars didn't need to be assembled out of so many parts, even though they had been since the days of Henry Ford. Instead, Tesla began experimenting with casting bigger and bigger pieces of the car and eventually succeeding, greatly reducing the need for assembly.
For insurers, though, I'm thinking the real benefit would come in more modest ways that track more closely with an anecdote McNeill told in a podcast with the Wall Street Journal. He talked about how much trouble Tesla had designing and manufacturing a part that was supposed to sit between a battery and the chassis. The problem became so important that Musk got personally involved and haunted the factory for weeks. Eventually, Musk and McNeill asked if the part was really necessary, and the battery people told them it had been mandated by the folks responsible for damping noise. When McNeill went to them, he was told that, no, the battery folks had mandated the part to minimize danger in the case of a battery fire. McNeill decided to track down the engineer who had signed the order requiring the part -- and learned he couldn't reach the person because he was a summer intern who no longer was at Tesla.
Insurers already question what they believe to be undue regulation, but I think they could benefit more broadly from asking employees across the business to question everything they're told to do, whether by someone inside the organization or outside it. Even if you just do this as a one- or two-month exercise, I'd bet you'll find you're doing lots of things just because they've always been done that way, not because they deliver any value.
Delete Every Possible Step in the Process
At Tesla, McNeill said in the podcast, he deleted several steps, and Musk asked whether he'd broken the process as a result and received some severe pushback. When McNeill said he hadn't, Musk told him he hadn't gone far enough. He needed to keep pushing until he not only got close to the bone but cut into the bone -- at which point, he should back off and find a sustainable approach.
McNeill said the rule of thumb was to only deliver what the customer directly paid for: the car. Customers didn't pay directly for manuals, for documentation, and so on, so Tesla would spend as little effort as possible in those areas.
Again, I don't think that approach would survive at an insurance company. Cut-until-you-break-something can happen in a manufacturing process, behind the scenes, but it didn't even work at the Department of Government Efficiency (DOGE), which Musk ran in the early days of the second Trump administration. Even with the slash-and-burn ethos of Trump 2.0 a year ago, Musk cut too deeply and caused problems both for those receiving government services and for Trump.
Still, insurers can suffer from a sort of data and process bloat. Given the industry's abundance of caution, it's easy to ask for more questions, to gather more data, and to require another guardrail in the process. Life insurers have shown that it's possible to do the same with less, given the success of fluidless underwriting, and other lines could surely scale back some requirements, too -- becoming more efficient while making customers happier.
Simplify, Accelerate, and Automate
I'm combining the last three parts of the Tesla algorithm because, at least for insurers, they all fit under one mandate: "Automate last."
McNeill said Tesla learned the value of these three steps when it was having so much trouble manufacturing the Model 3 that it was running out of cash and was in danger of bankruptcy. The company stopped running its highly automated manufacturing line, set up a big tent outside the factory and started making the cars by hand. Once management figured out the best process, it began speeding up. Only once they saw that they could run the process at speed did they start bringing in the machines that would automate it -- scrapping the entire production line that they'd set up before fully understanding what was needed.
"Automate last" fits with the approach the computer industry has taken for decades: Once a manual process is fully mapped out, it can move into software and then, when you're sure you have everything nailed down, you can hard-wire the work by moving it into the silicon.
That approach makes sense for insurers, too. When you see the possibilities of AI, for instance, you should map out a potential new process, implement it manually, speed it up -- and only then let the machines take over.
There are plenty of things about Musk's approach to business that I wouldn't recommend. For more than a decade now, I've been mocking his annual claims that he'll have millions of Teslas functioning as robotaxis, that he's going to colonize Mars (we won't even land someone on Mars in his lifetime), that he's about to unleash an army of humanoid robots, and so on. Those of us without his massive wealth would lose all credibility overnight if we pushed a similar sort of sci-fi dream. Insurance, as an industry built on trust, can't afford anything close to the wild claims that Musk makes routinely.
But I do think it's worth giving his algorithm serious consideration because it can reduce complexity and lead to effective automation. If nothing else, reading about the bold moves at Tesla might inspire some new thinking and resolve in the insurance industry.
Cheers,
Paul
P.S. "The Algorithm" reminds me of one of my favorite geek jokes:
Q. How do we know that Al Gore actually invented the internet?
A. It runs on Al-Gore-ithms.