Burned out on AI? Why IRL is making a comeback

Author

Shai Goldman
December 09, 2025

Happy Tuesday Brex Community,

I’m constantly online: reading tweets, seeing photos, videos, long form posts. And I increasingly ask myself “is this real, did a human create this?” Many of you are in the same boat, constantly analyzing if this digital medium that you engaged with was created using an LLM/AI (ChatGPT, Gemini, Claude, Grok, Ruway, Suno, etc). In 2026, the swing to IRL will be even more pronounced, as you know that the in person engagement is actually REAL and not AI 🙂 The in person conversation could fall flat or the person might not be the most candid, but at minimum you know this experience is 100% real and in this increasingly AI world, that is satisfying. My challenge to you all is to do more IRL engagement with your current and future customers in 2026 than you did in 2025.. Hop on a plane to close the deal, switch the meeting from zoom to an in person coffee if you are both local, bring your customers together for a nice meal, do a white boarding session in the office, head to that conference; surprisingly these things are still not the norm for many business people. IRL meetings do take more time in terms of travel, so it is easy to default to video/voice call for the sake of efficiency, however I find that the ROI is more impactful in IRL!

Have a great week and no, I didn’t use AI to write this 🙂 

Shai Goldman

Dec 10 - Brex Startup Holiday Party - SF - for Founders and VCs - apply here

Dec 18 - Consumer Holiday Bash w/ VHS Ventures - NYC - for Founders and VCs - apply here

Meet Jeff An and Wei-Wei Wu of Momentic, an AI-native test automation platform. They help engineering teams at companies like Xero, Notion, Bilt, Webflow, and Quora catch bugs earlier and ship faster while avoiding the maintenance overhead of traditional test suites.

Problem: When engineers build a website or an app, they constantly update it with new features. But every time they add something new, there is a risk they accidentally broke something old (like the "Login" button suddenly stopping working). Checking everything manually before every update is boring, slow, and expensive. Engineers hate doing it, and humans often make mistakes or miss things.

Solution: AI robot that does this checking for them. Instead of writing complex code to test the app, a human just tells Momentic in plain English: "Go to the homepage, log in as a user, and try to buy a t-shirt." AI Execution: Momentic’s AI understands those instructions, opens a web browser, and actually clicks through the site just like a real human would to make sure it works. Self-Healing: Usually, if an engineer changes the color or name of a button, automated tests break. Momentic is smart enough to "look" at the screen, find the button even if it changed slightly, and keep the test running.

Why it Matters: Speed: Companies can release updates much faster because they don't have to wait days for humans to manually test everything. Reliability: It catches bugs before you (the customer) see them. Simplicity: You don't need to be a coder to use it. Product managers or support staff can tell the AI what to test.

2 Interesting Posts This Week

Sign up here using referral code NEWSLETTER.

Trusted by 1 in 3 US startups, Brex is the financial stack you can bank on — from MVP to IPO. Get up to $6M FDIC insurance, no-fee corporate cards with high limits and top rewards, built-in spend management, seamless travel booking, and AI-powered bill pay.

💡 New customers unlock over $180K in partner perks, including $5K AWS, $4K Google Cloud, and $2.5K OpenAI credits. All free to use.