A friend who's been in the sales business for decades recently told me he was afraid that Artificial Intelligence (AI) would soon replace 80% of today's salespeople. I'd be worried, too, if AI hadn't turned out to be a technological dead end.

Despite the wide-spread belief that computers are getting "smarter," the day-to-day applications that use "AI" in business environments are so bad that they're laughable.

Take, for example, the automated customer telephone support lines that claim to be able to "understand complete sentences." Bullsh*t. Unless you're doing something obvious, like checking an account balance, it's like talking to an idiot.

And that's not surprising since those "voice support" applications combine two technologies that have never worked very well anyway: natural language processing and voice recognition.

Natural language processing is what websites use for those "free-form" queries that point you at the wrong FAQ answer. Voice recognition what corporate voice mail systems use to connect you with the wrong person.

Both natural language process and voice recognition are decades old and have only improved marginally since they were originally programmed, even though the computers they run on are 300 million times faster than they were in 1985.

So, then, what about the widely publicized "breakthroughs" like IBM's undefeatable chess program and Google's driverless car? Don't they show that computers are getting smarter and maybe smart enough to become salespeople? Well, no.

Any average chess player can easily defeat IBM's chess programs by insisting upon playing a variation, like Turkish Great Chess, which has pieces that for instance combine the moves of a knight and a queen.

Even the best chess program--unless reprogrammed with the alternate rules--would simply fail out the gate. By contrast, if you stuck a human grand master in front of a TGC board, he or she would immediately adapt and totally kick your butt.

Similarly, Google's driverless car isn't emulating the intelligence of a human driver. It only works because Google built an incredibly detailed map of the area where it drives, and even then even a dusting of snow would totally screw it up.

In other words, Google's "breakthrough" a big Roomba. And, like Roomba, it uses the same tired old algorithms that computer games use to move non-player characters around a virtual map. Ho-hum.

I pointed all this out to my friend and response he cited an article where some famous computer scientist predicted that within 20 years, computer science would create a "singularity" that would be able to think like a human brain.

However, "computer scientists" have been predicting that "singularity" since the 1950s and it's always something that's going to happen in 20 years. And it never does.

Here's why: neurons are astronomically more complicated than computer bits. Far from being a binary switch, neurons pass electrical wave patterns to each other influenced by a complicated set of chemical reactions.

If anything, a neuron is quantum computer, not a bit. And if that's the right analogy, a single human brain is more complicated than all the digital computers in the world today, combined!

Jobs biographer Walter Isaacson recently expressed his opinion that the "singularity" scenario remains unlikely and that it's the combination of human intelligence with computer processing--not "AI"--that's changing the world.

In Sales, that union of human and machine is already happening, big time. Sales technology (CRM, email marketing, etc.) is a multi-billion dollar business. What's not changing, though, is the complexity of human-to-human interactions.

According to recent research from LinkedIn, the demand for talented salespeople far outstrips the supply almost everywhere in the United States and Canada. So, AI robots replacing human salespeople? Not happening now and probably not ever.

So stop worrying and keep selling.