Reading today of Bigpond's plan to add an IVR in front of their tech support gave me the following business idea: What if, instead of having to wrestle with IVRs and/or wait on hold for customer service, you could have someone (or something) else do that for you and call you when they've reached a human?
There are two things I'm aware of that sort of do this today:
- Callback - some call centres allow you to leave a number so they can call you back when an operator is free, rather than have you wait on the line. This is a great feature and I'm not sure why more places don't do this.
- Gethuman - a listing of many IVRs, mainly in the US/Canada and what sequence of buttons you should press to get a human; note that it isn't so helpful for IVRs that require voice input
Building a system to fully automate this would not be so hard. You could work off a prioritised list of the most frequently used call centres - e.g. big banks, utilities, subscription services (PayTV etc), telcos, ISPs and so forth - experience shows its mainly larger companies that bother with IVRs, and implement them in the most annoying way possible so there's not that large a list. Work down each company working out the shortest path to a human, whether that's through voice input or PSTN tones, then simply record those.
Now the customer calls you (this needs some thought), you log their number using caller ID or through manual input from them and the call ends. Your service dials the customer support line, plays back your recorded sequence, and upon detecting a human voice you play a "excuse me the kettle's boiling" and dial the customer back, connecting them to the customer service rep.
I don't have a super elegant solution for the initial customer contact. One option is to outsource that to your own overseas callcentre who understand enough to know that the customer wants to contact "Bigpond" or "Telstra", or "NAB" or "Commonwealth Bank". Alternatively you could have a different phone number for each and move the mapping problem to the end user, or, (oh the shame), your own IVR with a selectable list. :)
The business model for this would basically be user pays. Given that automating this is entirely feasible I think a flat fee model is simple and works best. Maybe use a premium in-bound 1900 or SMS which bills the user $2-5 or something equally small - given the basically zero marginal cost of service I think going for a smaller price makes sense - it makes it cheap enough for anyone who becomes aware of the service to give it a try since its not much to lose, and a helluva lot better than sitting on hold for 20 minutes.
