If you ever talk to a tech investor, you’ll probably hear words like “unicorns”, “J Curve” and “Scale”. These terms all come from the concept of putting money behind a company so it can grow to a ridiculous size and the investors can make buckets of money. And one way to do this is to commoditise as much of the business as possible so the business can “hit scale”, i.e. service as many people as possible with as little cost (effort) as possible. And there’s no easier way to do that than using technology to automate tasks (if not whole jobs).
So, you may have noticed, it’s getting rarer and rarer to be able to interact with a human, especially one located onshore. That is because the masters of scale will use technology whenever possible to service you, and when not possible, the next option is to use a cheap human resource offshore. Onshore humans are considered to be expensive, but that is a false economy (I’ll get to that later).
It’s Only Cheaper if it Works
The benefit of all this automation to us consumers, (or so they say), is lower prices. The problem is the solutions don’t always correctly address the problems at hand. Often, the computer (which will interact with you via a software system, chat bot, or an interactive voice response system doesn’t always understand the nuances of what you want, or very often, even comprehend the problem being solved. That is because to solve a problem in a commoditised way, it needs to be systematised (I.e., been previously seen and addressed, usually by a human, with a system, which is then captured in software; this is known as “automation”). So, if your problem doesn’t neatly fit into the problem set that the system was designed to tackle, then you have a challenge to deal with. And often, things are set up in such a way that it’s extremely difficult to get out of the system so that a human can help you.
But won’t AI Fix It?
Now, you may have heard about how Artificial Intelligence (“AI”), specifically Machine Learning (“ML”), is going to be able to identify the problems that the humans haven’t previously. ML is great for finding patterns in data, whether they be what we’re specifically looking for, or exceptions. The problem is there’s no emotion in raw data. And Yes, AI has an area focused on “sentiment analysis”, but that is heavily reliant on the quality of the data and often misses things like cultural context and sarcasm. So, algorithms and large data sets (a.k.a. Big Data) won’t always get us the solution to our individual challenge; the commoditised version of this is known as “mass personalisation” (sounds like a contradiction to me).
How a Human Helped Me
So, let me tell you about an experience I had the other day – I had an issue with my internet connection, so I thought about whether or not I wanted to lose hours of my life and make a support call. After considering all my options I concluded I had no other choice. Given the size of my ISP, I assumed I would have to go through a myriad of choices on the interactive voice response system, before being put on hold for an hour and hopefully finally get to talk to a human. So, I dialed the number and switched on the kettle thinking I had plenty of time to kill before I had to engage with someone. Low and behold, my call was picked up by a human in less than a minute. What was even better was that human, Marco, solved my relatively complex problem in less than a handful of minutes and not only that, rewarded my long tenure as a customer by sending me out a brand-new modem that supports dual-band wifi, at no extra charge! So now I’m enjoying double speed wifi and a working system because a human made an intuitive call and decided to reward a long-standing customer. The result is I am now feeling particularly happy with TPG, who up until that point I had assumed would treat me like a number. Thank you to Marco at TPG.
Do it Right the First Time
So, using machines and lower-cost humans (who often work from a series of scripts) may create savings at certain times, but often they cause the exact opposite outcome. There is definitely a time and a place for it, but humans are also needed in servicing humans too. There’s no substitute for a knowledgeable, experienced human, who speaks the same language as you, (and as well as you do, ideally). Or put more simply, if something is done right the first time, it ends up being much cheaper than being corrected multiple times; and I don’t just mean in terms of direct costs. Think about the costs associated with lateness, re-work and most importantly, brand costs.
Ultimately, even if machines can behave like humans, I don’t want to be tricked into thinking I’m dealing with a human when I’m not, and nor do I want to create a “connection” with a machine. It’s important that only the commoditisable is automated, while the value-add is also catered for (by humans). To me, there’s plenty of jobs in the future job market for humans because, at crucial times, I Want To Talk to a Human!
Note: no machines were destroyed in the making of this article.