When flights stop, the first person to ask may not be you
It was 11 p.m. on Tuesday, and I was still tidying up a quote when a client sent me: “Is my flight delayed again?” My first reaction was to open several flight-tracking pages and check each booking one by one. With that many updates, I missed the client with the latest departure. A three-person photography team also had no idea that the airport had suspended all departures, so they had to contact more than a dozen customers individually. We freelancers who work remotely know that “We’re dealing with it too” anxiety all too well.
I built a small “exception alert” workflow
The idea wasn’t complicated: pull status updates from flight-tracking pages, keep only records tied to client itineraries, and whenever a delay, cancellation, or airport shutdown appeared, create a notification list and let AI turn the required explanation into a short message.
Xiaolin, who builds custom itineraries in Hangzhou, is already using this approach. After a severe convective storm in March 2025, she put 28 clients’ itineraries into a monitoring sheet and sent alerts only to travelers whose flights were affected, rather than broadcasting to everyone. AI only rewrote “Flight cancelled” into a polite, concise client message; a person still checked it before sending. The key wasn’t letting a machine decide what your clients need, but avoiding another round of refreshing the same information.
It doesn’t cost much to recreate, but define the boundaries first
My replication cost: roughly RMB 30–80 per month for tool subscriptions, plus four to six hours to build and troubleshoot it. The technical barrier is understanding spreadsheet filters and copy-pasting; no coding is required.
For a first attempt, I’d open the flight-status site I already use, search for “flight status” or “airport status,” and create a spreadsheet with each client’s name, flight number, date, and notification status. Then I’d find an AI tool that can read web pages or spreadsheet content and draft the notification. Some sites restrict automated reading. When that happens, I wouldn’t force an integration; I’d switch to checking twice a day at fixed times. Not everyone needs this setup. If I only handle one or two clients a day, manual checks may be easier. There’s no urgency to try it now.
Match the workflow to client volume
If I were just starting out with a handful of clients, I’d begin with a spreadsheet and put “check flight—record status—send notification” on my daily checklist until the routine felt natural. With one or two steady clients or more than ten itineraries a month, I’d add an automated alert workflow, but run it manually for a week first to catch false positives. Only while scaling—if flight questions repeatedly interrupt customer support—would I consider connecting an order system, standard message templates, and human review.
This tool isn’t for everyone. What I’m really buying isn’t more automation; it’s one fewer missed conversation and one fewer late-night round of status checks.