I have been working on one of the first “I.V.R.” (Interactive voice response, for those not familiar with the acronyms) back in 1997. It was a huge thing back then, especially as ASR and TTS (the natural language way of speaking with a machine) were really early technology and to say the least, a bit immature. I remember that we randomly called people from all over the country and recorded the calls for the ASR engine to get all the slang and accents.The embryo of today’s voice portal was in my hands and it looked awesome and cool and as an engineer, the potential seemed so high that we imagined these things would replace humans in just a few months!
So after a while we enthusiastically rolled out our first scripts: developed, tested, re-tested, documented and highly appreciated by the techies, only to find out that
…they were a complete disaster once in the hands of the end users.
They lacked some instructions, albeit the intuitive menus, some quick fallback to a nice sweet helping voice, the speech recognition was funny and bizarre and text to speech failed to be comprehensive, but most of all, we assumed that common users did prefer a machine to a human. I am still convinced that some of the errors we made back in the old days are still haunting us because guess what? People remember.
Now things have changed a lot, but again not so much from those years:
- We still develop scripts for the user to play with – sometimes we also provide some sort of independence for the users to develop their own dynamic script during the interaction. This still has lots of potential but is inexorably linked to the target users.
- Adoption rates largely vary according to the service, the target demographic and the goodness of the script. And I am not mentioning here those ivr trees that are specifically developed to confuse and deflect people from calling.
- ASR and TTS are still not widely accepted for many reasons that have nothing to do with the technology, which is even more awesome and amazing than ever (have you tried it?) : namely the end user willingness to accept the use of natural language in their game with an automated response thingy.
- Self-service is still somehow big: people love to find their own answers and browse safely on the web or on any other device that will guarantee some degree of anonymity: similar to the unparalleled satisfaction of mounting our own Ikea piece of furniture (for those of us who don’t end up with a weird library and a bunch of unused screws), we love proving to our self that we can be self-sufficient across different channels and fix our own problems
- A human representative is still the choice of election for those problems that are not in the Q&A, not in the FAQ, not in the knowledge base and not in any automated script and therefore require the human brain to solve it.
How are we doing with this? 18 years and still dealing with ivr scripts? There must be something new in our technological horizon, something that will lift us to the point where users will find 99% of their answers and the 1% will be answered by a super-human agent who is someone like Mr Wolf, only “solves problems”.
For now, I guess we’re far from that target and I really need to call my isp support to fix this upload issue… 😦
2 thoughts on “Diaries of the voice portal”
thanks Marco, you are perfectly right and ahead as usual 🙂
IVR, ASR…for how long will companies still sell and use this crap? Especially after so many years where Telcos or Utilities operators have used these only to deflect the vast majoirity of clients because belonging to one of these categories that marketers have decided not affecting the churn index! More and more people (even my mother or my stepfather, not to mention the young generations) are chatting through Facebook, Whatsapp, Telegram, Skype, Twitter, etc on multiple devices that comes handy (and cheap) in your pocket or at home. If on the other side there is a live person typing the answer (to be noted that it might not even be someone on a traditional call center) or a digital customer assistant (nice wording replacing the so innatural and cold virtual agent definition) it doesn’t matter as long as the answer is pertinent and accessible anytime through any device in a reasonable time. Yes, nowadays typing produce more effectiveness and less frustration than speaking to/with an IVR-ASR. Because is not the technology that is not good but it’s about the people in the companies that adopted (and sometimes even those that installed/configured the system) that made it so unfriendly, unusable and so easy to hate it. About contacting your isp for your upload issue I need to point you to what Albert Einstein said once: “The significant problems we have cannot be solved at the same level of thinking that created them.” 😉