(Customer) Service Oriented Architecture: Air France Fail
By stern on Apr 16, 2009
When I went to board the plane, though, Air France didn't find my ticket. This should have sent off loud klaxons for me, and should have initiated something of a "tie off the loose ends" set of transactions -- if I wasn't auto-checked in, then my bag was likely not going anywhere either. The gate agent re-assigned me into a seat, but the transaction died on the network vine. Somewhere around the equator, a flight attendant handed me a message saying I should contact baggage service in Johannesburg. Good news: they found my bag. Bad news: it was enjoying a vacation in France without me. As soon as AF realized that my ticket wasn't checked in, a series of transactions to find my bag, route it to the plane, and verify if it could be loaded within the departure window should have commenced without any whining or outside influence. Wasn't this the whole point of SOA?
While some analysts proclaim the death of SOA, the idea isn't bad, as long as the focus is on delivering some sort of result. What Air France needs is a mix of real-time request routing, SOA, and a focus on customer service.
Thirteen hours after landing, and 2 hours after my bag supposedly was supposed to arrive behind me, I still haven't heard from Air France. No idea where my bag is, or if or when I'll see it. They're doing their best to ensure I don't fly with them again, as it appears my luggage experience is far from unique.
If SOA is dead and social networking is alive, here's an idea: What if Air France's customer service people follow negative threads on Twitter or blogs? The "service" that matters here is customer service, whether automated or not, and human intervention in human problems frequently goes further than automated admissions of cluelessness.
Personally, I think it's the evil karma of wearing a suit that has come back to re-route the suits themselves away from me, but I don't have the tweets to prove it.