Remember meForgot password?
    Log in with Twitter

article imageOp-Ed: Can A.I. perform Extreme Personalization customer service? Maybe

By Paul Wallis     Apr 18, 2018 in Technology
Sydney - If “Extreme Personalization” sounds like the most offensive, riskiest buzz phrase ever, you’ll be pleased to know it comes with a tonnage of management science crap, too. The new idea is to make A.I. effective with customers, but…?
The theory of Extreme Personalization deserves a look. This is, ironically, designed to deliver what today’s incredibly lazy, backward customer services don’t – A one on one, customer-specific, productive service. The basic theory is that “markets are conversations”, and customer service, which sometimes includes meaningful conversations, is part of that.
The theory, however, is based on the equally difficult premise that anyone knows how to hold a meaningful conversation and more to the point, knows how to teach A.I. to do that. This is Customer Relations Management in machine language. It’s a big ask, in an environment where customer service is generally considered to be at an all-time low in terms of quality.
A.I. learns how to do things. There is some basis in believing that it can become much more than a chatbot, and hold an intelligent, productive conversation based on its learning skills. The bigger question is whether this cookie-cutter-don't-give-a-damn-we-hate-our-customers customer service management environment has any hope of understanding the value of that idea.
The realities of customer service
A bit of background – I wrote a book about customer service a few years ago, based on decades of experience with customers. I spent a lot of time at the face to face level, dealing with people who were given any old thing as information by human beings, or alleged human beings, and then spent ages trying to solve their problems based on that misinformation.
The usual problems, at least 50% of the time, were and are:
1. The person providing the information delivers incorrect or misleading information.
2. The receivers of information are misled, or misinterpret accurate information.
3. The stress and nuisances of information mishandling add a lot to the possible risks for both customer service providers and customers.
Now – Add an innocent A.I., trying to learn, to this mess. How good can the AI be, if quality of information is so slapdash?
Tech. Revolution has an interesting, if at times maddening, article on the many issues, which they have commendably and understandably managed to fit on one page. Tech Revolution points out that the original idea of digitally connecting people was overtaken by transaction-based technologies and business models. The result is the truly godawful customer service we see today.
I can add a bit to that – I did an interview for a job with a well-known global company which involved “wealth management”. I was told that I was a good salesman, but spent more than the allocated 3 minutes on the customer. Consider, if you will, that this customer could be worth millions, and I’m supposed to let them (and their millions), go because of an extra minute or so. As though these piddling KPIs and ridiculous rituals equate to millions of dollars. They don’t, by the way.
This mindless approach to customer service and communications has absolutely nothing to recommend it. The truth is that customer service has become yet another middle class processing job, where nobody does anything but follow the rules, however inappropriate.
Now consider teaching A.I. to be equally stupid and do non-business like this. It can happen, if “conversations” are limited by absurdities like this. It’d be better to let AI do what it’s supposed to do, have a conversation and simply deliver the required services.
I’m not going to regurgitate the Tech Revolution article. It should be read, preferably with a nice cup of something and no sharp objects around for customer service experts. It reads like Pandora’s Box, a risk management horror story with technologies added.
Will it work? It might.
I’m not at all against using A.I. for routine business, simple questions, or basic transactions. Actually, it’s a very good idea, provided it can really do the job. That said, I’m totally against it for any risk of losing customers, or creating legal death traps for businesses. How is A.I. supposed to know the basic legal obligations of suppliers to consumers, for example? What about statutes, contracts, faulty goods, provision of services problems, etc.?
Good AI probably could have a full bank of consumer law built in, and make good recommendations. The problem is that learning that volume of information and correctly applying it is tough enough for lawyers, courts, and consumers, let alone unsuspecting A.I..
I can see a raft of legal train wrecks if this idea isn’t rigorously, meaning almost at OCD level, got 100% right. As for Extreme Personalization – My advice as someone in the business is to lose that expression, ASAP, because extremely personally is exactly how customers will react to problems with AI systems.
The A.I. is coming to customer service, ready or not. Many corporations are currently either marketing something called A.I. customer service or researching it. You’ll be talking to the software soon enough. I wish you, and it, good, meaningful conversations and many of them. It's obvious that this tech will become standard, soon enough, and it will have to work, and work well.
Or the future is well and truly screwed, and screwed mightily.
This opinion article was written by an independent writer. The opinions and views expressed herein are those of the author and are not necessarily intended to reflect those of
More about Artificial Intelligence customer service, markets are conversations, Customer service, customer service communications, customer relations managemetn
Latest News
Top News