I mentioned earlier about a (less) frustrating way to deal with customer service bots. The chatbots have evolved quite a bit that some times friends tell me “Hey I feel I am talking to a bot but I am not exactly sure”.

So they ask me “How do I tell for sure?”

Well, there are lots of tips online, like this one, which are clearly not written by engineers who develop chat bots. For us (who have been developing bots), the point of detection would be targeted at areas where humans have not overcome while developing bot algorithms.

For example: bots are (still) not very good at understanding context, as humans communications in a less defined way. The algorithms have evolved to better understand and define the context – you probably have that experience with Siri or Alexa.

However, the pretext is that you know Siri or Alexa is a bot. , and you are not using it for a particular enquiry/transactions, which means you can answer questions as random as you want. Also, unless you are in China, you would be more likely typing to a customer service than calling.

When you talk to a customer service, you want to get a particular issue resolved/enquiry answered, and you probably do not want to offend (or make yourself sound silly) if the other side is indeed human.

Easy detection

One way to do it, in this context, is to breakdown your sentences in a way that differs from usual abbreviations or short cuts people use. Examples:

“I wa, to disz abut tkt for nxt Wsday”

“Und efter ze fifz yer, ve vil al be speking German like zey vunted in ze forst plas.” (you probably know this joke).”

“I cdnuolt bleveiee taht I cloud aulacity uesdnatnrd waht I was rdanieg. The phaonmneal pweor of the hmuan mnid, aoccdrnig to a rscheearch at Cmabrigde Uinervtisy, it dseno’t mtaetr in waht oerdr the ltteres in a wrod are, the olny iproamtnt tihng is taht the frsit and lsat ltteer be in the rghit pclae. The rset can be a taotl mses and you can sitll raed it whotuit a pboerlm. Tihs is bcuseae the huamn mnid deos not raed ervy lteter by istlef, but the wrod as a wlohe. Azanmig huh? yaeh, and I awlyas tghuhot slpeling was ipmorantt!”

You probably would have no problem reading the above – machines will find it difficult.

High context language 

In Chinese it is even more difficult for the machine to understand the context because words are not phonetic, and combinations (or sequence) of different characters can mean very different things.

Things like “中国队大败美国队“ can mean “The Chinese team has beat the US team” or “The Chinese team has lost to the US team”. You really need high context to understand – and this is more challenging for machines which are learning based on certain logic.

However, there are some companies who are determined to provide poor or no customer service (no frills). No matter whether they are using bots of human, the result will be the same.

Thanks for reading The Low Down (TLD), the blog by the team at Momentum Works. Got a different perspective or have a burning opinion to share? Let us know at [email protected].