COMMENT
Then yesterday a statement landed in my inbox. SpinVox "works with" five call centres, it said. But there was no word on where they are located, or the total number of staff employed. The company would only say that the "few hundred agents" that are now required per market have been cut down from "the thousands" per market when the business started.
According to a spokesman, SpinVox operates in eight markets so we can conservatively guess the call centre staff payroll is somewhere in the region of 2,400 seats.
What about the proportion of calls handled by The Brain vs those going via the lughole? The statement makes some grand claims - saying the tech needs "two per cent of the [human] input" it needed two years ago; and that it can apparently "predict more than 99 per cent of what most people speaking in English or Spanish will say next". Who knew people were so predictable?
But it's hard to shake the feeling that I still haven't got the info I asked for. 'Ninety-nine per cent of most people' sounds like a lot but it can't be measured in any meaningful way - does "most people" mean 80 per cent? Seventy per cent? Or just a majority (51 per cent)?
And as for the "two per cent of the input" statement - all we glean from that disclosure is their system is better than it used to be, and for all we know The Brain's first week of operation saw it gurgling a string of consonants like a newborn babe.
Read this
silicon.com Skills Survey 2009
Click here to take our survey - and be entered into our draw to win an iPod Touch
If there's any lesson here it's that it pays to be clear - and not just when you're leaving a voicemail message. Businesses must as clear and candid as they can be without compromising 'commercial sensitivity'. Misunderstandings all too easily breed mistrust: something no business wants to engender in its customers.
SpinVox was fairly clear about human involvement in its system but perhaps not entirely transparent by the warts-and-all standards of a public that poured over every last dotted i and crossed t of MPs' expenses. And I would still like a straight answer on the proportion of voicemails that need a human ear to give up their secrets.
A quick straw poll of SpinVox users on Twitter suggests they are actually an easygoing bunch who don't mind if their calls are being transcribed the old fashioned way, by ear - just so long as something intelligible lands in their inbox at the end of the day.
But SpinVox investors and shareholders have a real reason to care - long-term profitability for this start-up from the class of 2003 is likely to hinge on the quality of its technology, and the quantity of its call centre staff. So let's hope their ears at least are party to what goes on in the boardroom, instead of the transmogrified discussions about the discussions.
And for the record, I'd wager the Welsh surname of Cellan-Jones is a reference to a group of Joneses who originally hailed from the mid-Wales town of Cellan, near Lampeter. SpinVox, are you listening?
Comments
There are 2 comments. Join the discussion
1. anonymous
When Daniel Doulton (is he D2 incarnate?) claims that "our automation has taken over 98% of the task" how does he define "task".
Let's say that a complete SpinVox service involves 50 steps, from receiving a message, to transcoding the audio, to checking for blank (slamdown) messages, to logging it, to backing it up, prior to conversion, and then after the conversion, there might be steps of formatting the message into an SMS, and formatting for email, and adding any advertising or marketing messages, steps for delivering the message, and then reporting and billing, etc.
So if only one of these hypothetical 50 steps was the conversion of the speech to text, and that conversion was done 90 per cent by humans, would that mean that the system is 98 per cent automated?
Just askin.
2. Jonathan Present
WHO PUT THE SPIN IN SPINVOX???
I believe that this company has received 100 Million in Venture Capital to date. Perhaps that gives them some latitude to get the job done, one way or another. However, they would require sophisticated speech algorithms to process the information in the manner they claim to. Do they have any Intellectual Property, Patents?