This edition is all about United Airlines’ new app: yay or nay? Financial platform Ramp’s prediction for the job market, the decline of Sweetgreen, and the thing I wish we won’t be doing in 2026.
Welcome,
This edition is all about United Airlines’ new app: yay or nay? Financial platform Ramp’s prediction for the job market, the decline of Sweetgreen, and the thing I wish we won’t be doing in 2026.
United Airlines knows something we don’t?
Remember the 2010s? Microbreweries and IPAs, MGMT playing on university campuses, hipster was as much a desire as a despise. It was also the time when every single company decided it wanted to build an app. Whatever you needed, there was an app for that.
Still to this day it seems crazy because an app costs a lot of money to build, and on top of it, companies were spending a lot of money to make you download it. App downloads were one of the main marketing KPIs at Accor, and the number one reason why we would never get 100 percent of our bonus.
You would hope that today, in the current economy and with the pace of technology, this kind of erratic behaviour would stop. That’s when United Airlines said, “hold my beer.”
They just release a new look and feel for the bag tracking, a progress bar for boarding, for whom do they build this? On my last United flight, which randomly happened to be last week, I only opened the app at security, to show my boarding pass. And I’m not the only one: according to United 85 percent of users only open the app the same day of their flight.
The problem isn’t just that the app is unnecessary, it’s that United is solving yesterday’s problem. Instead of using it, I asked Comet (Perplexity browser) to do my check-in and to summarize all info from my emails. It was still a little slow but it worked fine.
As LLMs become the OS of our lives, many previous workflows are becoming obsolete. Users don’t need a prettier app interface; they need their travel managed across every touchpoint without thinking about it. The goal is not to rebuild workflows one-for-one with AI, but to rethink entirely our way of operating. The faster we accept it, the faster we can move.
In their 2026 prediction, Ramp Economic Lab predicts that a significant share of customer service jobs is going to be automated by AI. It’s not going to be only in customer service, in many hospitality jobs too. The first I see automated is the night shift. Having one person, usually a student seated behind a desk for eight hours just to complete one or two late check-ins is not a viable solution. Airbnb property managers as well are going to rely more and more on AI for their guest interactions.
It was already to a minimum with lockbox in the street to make check-in self-serve, and a team of virtual assistants based in Asia to answer basic guest questions. Tomorrow AI will replace the VA.
But here’s the part that gets overlooked in all the automation doomposting: What I think we are not talking about enough is how these jobs are going to evolve, not disappear. Mark Cuban went on TBPN to explain why the next big thing for entry-level jobs is not to do the thing anymore but to implement AI. Think about it, someone needs to set up these systems, fix them when they break, and make sure they’re helping customers instead of frustrating them. As AGI is not near in sight, AI needs context and guardrails to operate effectively, and as the technology is new, this jobs are not going to be taken by senior level.
When the internet arrived, we didn’t stop using travel agents, we just started using online travel agents. AI automation follows this pattern.
After 19 years, one of Sweetgreen’s cofounders just walked away. 2025 was not an easy year for the founding team, and Nathaniel Ru’s departure was long overdue. Sweetgreen held the headlines this year and rarely for good reasons:
-
The $1 billion robot kitchen bet that went nowhere
-
A loyalty program collapse
-
The “potato debacle”
-
The “slop bowl” backlash, now cited as a recession indicator
-
A brand identity crisis spiraling in real-time
A more subtle signal for me was when back in March, Joe Hovde Hovde wrote his now-famous A tale of two salads, explaining how he switched his daily bowl from Sweetgreen to Cava without even a try from them to get him back. For a company who self-proclaimed themselves “a tech company before a salad company” it’s a missteps… any tech company knows the importance of growth levers and keeping your best customers inside.
Nonetheless, I believe Sweetgreen has the best merch game and accomplished something extraordinary, so kudos to their founding team.
2025 Outs
I am not inspired enough to do predictions and trends for 2026, but in the pure tradition of French people complaining, I can tell you the five things I want to see out before 2026:
-
Listening bar + Rey chair grape red from Hay: Seriously, if you are buying large speakers for your restaurant past September 2025, stop it and try to buy a personality instead.
-
Unreasonable Hospitality: We get it, you watched The Bear. But dropping the hot dog story from Will Guidara every single time you want to talk about guest experience is boring. Instead, go do a shift in your local pizza joint and you’ll have better stories.
-
Tip screen: You cannot mention 2025 without mentioning tip screen fatigue. I am not saying we should remove tips in the US, but I found this tip jar article inspiring: Bring back the tip jar, and people carrying cash!
-
Sharing plate concept: “Everything is to share, and we recommend three to four plates per person”… Joey doesn’t share food!
-
Members club: We already have great members clubs, we don’t need more. Especially one who is so healthy it doesn’t serve alcohol… instead, let’s all make sure Soho House becomes the darling it once was.
Happy holidays, don’t forget to water your Christmas tree (I was today years old when I learn this.)
