The future of text analytics isn’t artificial intelligence. It’s augmented intelligence.
It seems inevitable that, in the next few years, text analytics software will be able to understand your customers for you. It will be able to read and listen to every interaction, analyse the literal and emotional content and automatically trigger any necessary next steps. All you’ll have to do is sit back and watch your NPS scores rocket.
Or maybe not.
Here’s the problem (and the solution): Words don’t mean anything.
I know, it’s a strange thing for the world’s best writing agency to say, but hear me out.
Text analytics software analyses the strings of characters that make up text, and the collections of strings that form sentences. Here’s what that might look like:
116 104 101 032 116 104 105 110 103 032 121 111 117 032 099 097 114 101 032 097 098 111 117 116 032 109 111 115 116
People can teach the software to find particular patterns in this data, and you can even use meta patterns to infer emotions and concepts from it. Good software can dig out the most salient patterns from vast amounts of data that humans simply don’t have time to process.
But here’s the thing: that string of numbers above? Here’s what it is, in English:
“the thing that you care about most”
What does that mean to you? Whatever it is, it doesn’t mean the same thing to me.
And, in order to appreciate the range of meanings that those words might inspire in people, you have to experience the world for yourself. You have to know what it feels like to see something, to touch something, to move about in the physical world and to live in a particular time and place. When we read, we don’t just associate words with other words. We associate them with our life experiences.
In other words: we create meaning. It’s not in a string of numbers or letters.
Text analytics software by itself can’t do this. It can take enormous amounts of unstructured data, digest it, and present it in a way that gives us a chance to make sense of it. But that’s where it stops. It can’t truly understand the subtleties of linguistic interpretation or the human condition, so it can’t truly understand your customers. Text analytics software can augment our intelligence, but it can’t replace it.
Yes, artificial intelligence will one day be able to fulfil all the promises of text analytics, and a great deal more. But it will require a holistic understanding of the world, not just cold strings of ASCII characters. For now, let’s use it for what it’s good at, and not ignore the most powerful processing unit on the planet that already exists in our own heads.
At Verbal Identity, we’ve spent years training our brains to recognise the salient patterns and the subtleties of the world around us, and around you. So, if you’d like us to augment your analytics capabilities with a combination of software, linguistics and psychology, email Chris. Alternatively, you could always wait 20 years for the singularity.