AI – closing three UK call centres

and laying off 2,000 support

staff, notes Teresa Cottam, chief

analyst, Omnisperience, author of

Humanizing AI: putting people at

the heart of customer experience

from the TM Forum.

The claim is that customers

don’ t want to call customer service

agents any longer but prefer to

use its app and website. Read

beyond the headline and the real

reasons soon becomes apparent:

battling heavy losses it needs to

balance its books and is turning

to AI-empowered chatbots and

overseas-based agents to reduce

its OPEX.

So, what’ s the truth? Is the

call centre on its deathbed? Is the

future machine-based chat with

algorithms to cosplay caring?

It’ s certainly true that

operators are excited about the

potential of AI to transform

customer support. When

members were asked by the

TM Forum where AI / machine

learning had the most potential

to positively impact the telecoms

business, 87 % said in the

customer experience / customer

support domains. Which is hardly

surprising given that these are

both vital and very expensive. Add

to this the fact that customers

are never happy and increasingly

want more – more availability,

more personalisation, more

support for far more things.

But it turns out the devil is in

the detail. Sky’ s decision came

after a 10,000 customer survey

revealed that nine in 10 wanted to

be able to contact the company in

more ways than over the phone.

But the key here is customer

convenience. They didn’ t say they

didn’ t want to talk to customer

support.

What customers actually

want is instant, 24 / 7 support

in their channel of choice.

They want proactivity from

their operators and they want support to be effortless on their part – no queuing, no repeating information, and no being tied to a device that might not be convenient for them.

The challenge of delivering fast, efficient support over an ever-wider range of channels for an increasingly complex range of services means that operators are currently doing two main things. Firstly, they’ re automating their response to customer enquiries as much as possible( with world-class operators already automating up to 75 %) and, secondly, they’ re giving agents smarter, AI-based tools, which is transforming their efficiency.

I could tell you that AI is going to be a big trend in customer support, but that is self-evident. Instead, my prediction is that we’ re on the threshold of a resurgence in voice. Don’ t be fooled by the old thinking that everyone prefers text. Text ties you to a device and is slow. Voice

is a far more intuitive and efficient means of communication because, after all, we’ ve been perfecting it for millennia. The other major advantage is it’ s hands-free, which is important in a range of

scenarios( from smart cars, to factories to kitchens).

Which is why you’ re going to see the emergence of sophisticated voicebots which are available over an expanded range of channels( such as smart speakers and wearables).

Imagine you’ re having problems streaming Conclave, the most natural thing in the world is to ask your preferred smart assistant( Siri, Alexa, Google Assistant), what the problem is rather than go searching for your phone or trying to fire up a laptop. Back-end integrations mean that behind the scenes automated technical support services can analyse if there’ s a problem in your area, if your FWA unit is attached to a congested cell, or whether your router needs rebooting. Eight out of ten times the assistant will be able to reboot your router, advise that engineers are on site fixing the problem, or walk you through reinstalling your firmware.

But if the problem really is too complex to resolve, then the smart assistant should be able to seamlessly transfer you to a human without you needing to change device. The agent will instantly see your verified ID and all your support history before them- what the problem is, what’ s been ruled out, and whether this has happened before- along with recommended next-best actions. You don’ t have to queue, repeat yourself or go through the same steps again. You also don’ t have to change device or change from voice to text- leaving you with more time to get back to what you really want to do … such as watching Ralph Fiennes in a cassock.

Meanwhile, a little more AI wizardry will mean your support call is documented automatically – removing a frustrating task from the agent – and any insights shared. A closed loop that’ s important to incrementally improve performance. Postcall the agent who helped you will also get feedback from the screen, with offers of training modules to improve their performance further.

While it’ s certain that AI will play an increasingly important role in customer support, humans will still be an essential and integral element to customer support workflows. After all, someone needs to tune the model and keep it tuned. The most complex or novel problems will need to be solved by a human because machines are not yet creative. And humans need to set business priorities. As intelligent helpers, AI and ML will support the provision of effortless, personalised and customercentric care, with their ability to spot patterns and join up the dots helping us move from reacting to problems to preventing them.