Why we started Paloma

In August I visited one of our customers and ended up having a heart-to-heart with their CFO. He shared how the company sometimes feels like the houses in Venice that stand on wooden stakes driven into the water. The stakes, he said, are the handful of their core people the company is sitting on, as the water slowly rises. Sometimes he feels like he’s running against time to turn it from an operation that runs on people into one that runs on systems.

That image stayed with me, because I’ve seen it in every company like this I’ve been to, including the one I used to work at before starting Paloma. There’s a core group of people who really care, and the operation is running on their backs. They’re always overworked and always behind. Almost always, the owner is one of them. They’d rather not take any time off at all, because they’d end up working through it anyway, or at the very least be on call, and the blowback when they got back would be bad enough to make them regret ever taking it.

You can see it in how the work gets done. Orders come in over WhatsApp and phone calls, and the rules for which customer gets preference when push comes to shove are in somebody’s head.

People end up there because the work had to get done and the tools they were given failed them. The ERP is all governance and no real-life flexibility, so when it can’t keep up with how the business really runs, people go and run the work in spreadsheets. And the spreadsheet works, but it’s never a living source of truth. It isn’t connected to anything, it’s prone to error, and you end up doing things in super hacky ways like color coding cells, or running most of the operation out of that one sheet with eight macros that takes an hour to load, grows every month and is a ticking time bomb.

When we start talking to a company like this about fixing it, the first reaction I usually hear is “I don’t know how this is ever going to work,” or my favorite, “yeah but it wouldn’t work for us.”

I’m actually a pessimist by default, and in most cases my confidence is a lagging indicator. So why do I feel so much conviction that this is possible? Because my co-founder Alex and I have lived through this too and came out the other side.

In the earliest days of Deel, a company that today employs and pays people in 150+ countries, I ran payroll with a VLOOKUP on a spreadsheet that got so big it eventually timed out (afterwards, Alex changed that to an INDEX(MATCH) and that carried me for another two months, ha!). Our operations were so stressful that one time I had an actual panic attack over early Christmas payouts. I was firefighting fourteen hours a day, and in the middle of it, my first performance review said I wasn’t being strategic enough. I couldn’t be. You can’t think about strategy when 110% of your day goes to holding the ceiling from falling down.

What got us out of that rut was learning to untangle a complicated process by first understanding the core problem we were trying to solve. It sounds stupidly simple, but you’d be amazed at how much of everything becomes solvable if you take the time to first lay it all out in a flow diagram. So, we’d start there, imagine what an operations manager’s life could look like in an ideal world, and then design outwards from that. We could afford to work that way because there were entire software development teams dedicated to doing this work.

My husband Rashid runs an agriculture company, and when I look at his operation, I can see that at its core it’s not that different from ours at Deel. It’s a different sector, but ops is ops, and I keep thinking, here’s how we solved this before. He’s just as smart as the people I know in San Francisco, and he spends his days dealing with minutiae at an altitude they’d never have to. There are a thousand stones he knows he’d find something under, so he’d rather not turn them. We joke that at his job no news is good news, because most news is bad news.

And the stakes are higher than ours ever were. At Deel, one of the worst mistakes we did was accidentally refund customers their deposits twice. It was embarrassing, but we called every customer, asked for the money back, and it was fine. The worst thing that can happen to Rashid is finding out that a field is dead at the same time as his customer does. That field is written off for the season, the loss comes straight out of his bottom line, and he gets one season a year.

In the rare cases where he gets to do something strategic, it usually pays off. He’s a good businessman. Imagine if he had the bandwidth to make ten bets a year instead of one. He poured his life into this company and took an enormous career risk to build it. And after all of that, there’s still a precariousness to it that sets the undertone for everything else, partly because he has no end-to-end visibility and partly because so much of the work is crisis management.

When everything depends on you, the weight gets heavy enough that you almost wish something serious but not fatal would happen, just so you could take a few months off. Somewhere along the way he accepted that this is what his life is going to be now. Why should it be?

Alex has watched his brother run his own business the same way. Neither of us thinks people this capable should be spending their days putting out fires. Take the firefighting away and the same people do the same job at a much more strategic altitude. That’s why we started Paloma. We want their companies to run the way a tech company does, on systems instead of on the backs of a handful of people.

Most real-economy companies can’t afford a team of product managers, software engineers and designers to rethink each and every process from first principles. And even if they could, it’s not what they do. Taking everything a company does, making sense of it as a whole, and then figuring out what this should look like with what AI now makes possible is a hard problem to solve. At Paloma, we took what we did at Deel and dialed it into a method we can apply to most real world operations.

In the past, this level of custom systems design was only possible as a years-long, multi-million consulting engagement. But with AI, what took us months in the early days of Deel now takes days. That means we can deliver it repeatably, at a price that works for the kinds of customers we want to serve. The nice thing about AI is that there’s now much less downside to starting with something and iterating as you go. So it no longer has to be an all-or-nothing engagement that you commit to, like an ERP rollout you regret two years down the road. You can start one workflow at a time and fold more in over time. It’s an evolution, but it’s entirely possible.

I know a lot of people in the industries we serve look at AI and see either a pipe dream or something that doesn’t apply to them beyond ChatGPT. But when we look at an operation like this, it takes us back to our early days at Deel. We recognize the same firefighting, and also the same urgency and the desire to do good work, and we feel for the people in it. Serving them is important to us, and we get a lot of joy and pride out of making life better for people who care about doing a good job.

So if you’re one of the people your company is standing on, this is for you. You deserve the same capabilities and the same agility as any tech company. And you should want this for yourself. Not because you’re worried you’ll fall behind if you don’t adopt AI, not because companies like ours scared you into thinking you will, but because you want better for your own business. It’s possible, and you deserve it.