Proudly sponsored by

How artificial intelligence is growing in the workplace

Machine learning, task automation and robotics are increasingly being used in business.
Incorporating AI into the workplace could improve dynamics but the trick is to use it to help businesses grow and work well with humans. Picture: Shutterstock

So far the effects of artificial intelligence (AI) have been slow to reveal themselves in businesses in South Africa but the scale of the oncoming change is starting to become apparent overseas. 

AI’s influence is growing in the workplace and will bring substantial change to South African offices in the next few years as machine learning, task automation and robotics are increasingly used in business.

The ability of computers to learn, rather than be programmed, puts a wide range of complex roles within reach of automation for the first time.

Bots and virtual assistants

As machine-learning trained systems gain the ability to understand speech and language, so the prospect of automated chatbots is becoming a reality.

One example is UK electronics retailer Dixons Carphone, which used the Microsoft Bot Framework and Microsoft Cognitive Services to create a conversational bot. 

Google demonstrated the potential of chatbots last year with its demo of its Duplex system. Duplex rang up businesses such as a restaurant and a hairdressers booking an appointment while sounding and behaving enough like a human.

Household names are also muscling into the area of creating a virtual assistant for the enterprise space like Amazon’s Alexa for Business. With many AI-assisted technologies, the aim of using chatbots and virtual assistants appears to be either making existing employees more effective or replacing manual roles. 

Workplace sensor technology and analytics

Huge amounts of data can now be collected from inexpensive sensors applied to smart decisions. For example, South African workplace sensing technology company MakeSense allows businesses to accurately assess just how much of their workplaces they actually use, likely saving a lot of money in the process. 

It works by placing small sensors around the office which analyses peoples’ movement.

Workspace occupancy sensing technology helps businesses understand how desks, meeting rooms and break out spaces are used in extraordinary detail. For example on average 40% of people don’t turn up to meetings so many meeting room are probably too big and are wasted space and cost.

Machine vision in the workplace

Machine vision is an area of AI that could allow the automation of many manual roles that until recently would have been considered too complex for a computer system to handle.

A case is point is Amazon Go, a grocery store where shoppers just pick up what they want and walk out of the shop with their goods. The system works by using cameras dotted throughout the store to track what each shopper picks up. The shopper is charged when they leave, via an Amazon app on their smartphone.

Robots in the workplace

Robots are nothing new in the workplace, having been a fixture in car manufacturing plants for decades.

But what’s different today is that robots are beginning to be used for less repetitive and predictable tasks. Robots can increasingly cope with a greater deal of uncertainty in their environment, broadening the tasks they can take on and opening the possibility of working more closely alongside humans. Amazon again is leading the way in using robots to improve efficiency inside its warehouses. Its knee-high warehouse robots carry products to human pickers who select items to be sent out. 

Robotic Process Automation

Back office tasks like data entry, accounting, human resources and supply-chain management are full of repetitive and semi-predictable tasks. 

Increasingly, Robotic Process Automation (RPA) software is used to capture the rules that govern how people process transactions, manipulate data and send data to and from computer systems, in an attempt to then use those rules to build an automated platform that can perform those roles. 

Change is therefore coming to all workspaces all around the world; the trick will be getting AI to help business grow and work well with humans.  

Isla Galloway-Gaul, managing director of Inspiration Office. 


Sort by:
  • Oldest first
  • Newest first
  • Top voted

You must be signed in and an Insider Gold subscriber to comment.


“The ability of computers to learn”

Where do you you get THAT from?

Making a statement like that as if it is a given fact,
especially against the backdrop of the two Boeing-747-Max-8-Crashes,
when computers actually REFUSE to follow instructions!,

sounds crazy to me.

The Boeing program was designed and written by humans. On top of that, there were poor and even outright bad business and risk-management decisions made by human executives.

In Boeing’s type of business the consequences can be horrifically public and fatal. In most other businesses similar consequences can be just awful also, but without the embarrassing public display of “fatalities”. Think Steinhoff, the recent collapse of large construction companies, and the dismal ineffectiveness of the criminal justice system.

Seems that the human integrity aspect of human decision-making could do with some auditing by an unemotional AI?

Make no mistake, AI is coming and will impact society in a BIG way.

For example, even at this very early stage of development, autonomous self-driving cars are already showing results that suggest they are SAFER than the humans on the road.

The business decision regarding AI is very simple.

Businesses that do not embrace AI will be DESTROYED by their competitors that do.

This is lot of rubbish. Computers cannot be creative or think. They can only do as programmed and do it quickly.A simple flying bug is more complicated than the most sophisticated computers on earth.

LULU Alert, I just wish more people have your level of common sense.

Jonn Oxx, your first sentence, “The Boeing program was designed and written by humans”, made me smile.

Any program systems written by non-humans in operation anywhere? 🙂

You’ve lost some credibility talking about Boeing 747-Max-8….


This software is a perfect example of AI which CANNOT learn!

No MJ, it’s a perfect example of a badly formulated algorithm written by HUMANS.

Despite your mocking ignorance, there actually is a thriving field of development and R&D of “self-learning software”.

Computer code of large systems like Boeing’s or Windows is now so complex that it is virtually impossible for humans to confidently provide a 100% reliable QA on the safety or efficacy of it.

For this very reason there is ENORMOUS effort being poured into programming to write it’s own “self-learning” code.

Are we reliably there yet? Nope!

Is there still scope for human bad behaviour and AI mistakes? Oh yeah. And then some!

MJ Stellenbosch: I think Vanman is speaking to Jonnoxx – not to you.

Could be, Andries, but it seems that he was referring to mj’s incorrect reference to a Max 747 for the crash disaster

There’s no such thing. It’s the 737.

Now if mj had checked in with the commonly available AI Google Assistant (sies! ) before posting fake facts…


@mj Stellenbosch If a algorithm is figuring out on it’s own which parameters to fit to some dataset, it is learning on its own.

No human told the algo to choose a certain variable/value. It’s called unsupervised learning.

How does your email know how to decide which email is spam? There’s definitely not a person at google checking your incoming emails for spam, nor did you program the spam filter. It’s a algo “learning” by itself what each individual considers as spam.

Yes, humans still decide on the algo for now, but they definitely don’t train it.

Swanson, you say “nor did you program the spam filter” === Really?

Then “learning by itself what each individual considers as spam” == HOW?

Actually, my Google Profile stores the relevant information when I select an email as spam, and test this against future emails.

There is no self-learning involved.

@mj stellenbosch “The ability of computers to learn”

Yes Computers learn by storing your behaviour. Sites like Google, Facebook know you like sport or planning a trip; when you like sites and search sites. Using this information they will bombard you with ads relating to cheap airline tickets and restaurants etc. This is just an example of information collection and building a profile on you.

MUKS, you can call it learning if you want to,
but it is misleading.

It is still simple programming!

Do some research on machine learning and neural networks, and you will see that machines, (or computers) can do a kind of learning, predominantly based on logic and pattern recognition. The benefit is that they can do so on very large data sets without mistakes, (depending on the quality of the developers) and at very high speeds.

Have a look at the progress made in AI and agriculture as an example. There are now greenhouses that can produce 30 times the amount of food per hectare as traditional outdoor farming methods, using only two robots. The learning takes place by feeding the programs thousands of images of healthy vs unhealthy plants at each stage of growth, and so over time the program “learns” what is healthy or not, and so in real time can identify issues and proactively deal with them. They don’t get tired, they don’t get sick, or come to work with a babbelas, or go on strike, steal from their employer etc.

In terms of AI replacing human thinking and creativity, I believe, will require a paradigm shift in the way that computers work, i.e. a huge breakthrough in quantum computing or something similar, but until then, there is nothing that compares to the human brain in terms of creativity and intuition. But for repetitive tasks, admin and data analysis etc, AI will be a game changer.

However, for all the amazing aspects of the human brain, mankind as a whole is capable of some astonishingly stupid things!

Two thumbs up for this comment, Drachir!

Both Lulu and MJ could do with a bit of AI help on this subject.

Start with asking Google Assistant, guys.

“In terms of AI replacing human thinking and creativity, I believe, will require a paradigm shift in the way that computers work, i.e. a huge breakthrough in quantum computing or something similar”

This is just a dream, used to scare normal people! (non-programmers)

What you want to believe actually, is that quantum computers will become SO fast, SO big, SO powerful, they will obtain a consciousness somehow … …

To EVEN think that AI can replace human thinking and creativity, is the first sign of falling for fake news.

Computers are good in doing repetitive tasks and mundane tasks, and will not make a mistake as long as their programmer does not make a mistake.

Computers makes mistakes because of the programmers, and in THAT way, they can kill us, as stated in my first post regarding the 747 Max 8, which by the way, was just a teaser-comment … … …

Not really rubbish maybe the author could have picked better words. I’ve got a fair bit of experience with ML(Machine Learning). Look at this example that would hopefully illustrate: learning vs being “programmed/rule-based”:

Imagine an Alien comes down to earth today and for some reason he keeps getting puppies(dogs) and kittens(cats) mixed up(visually). You tried to give him a rule-set for identifying cats like:

-Cats have pointy years
-Sharp claws
-a tail
-Sharp teeth
-furry coat
-4 legs

But these “rules” works equally well for puppies also, a rule-based-programming-paradigm system (as most computer programs are today) is really not a great way to solve this problem: Cats Vs Dogs (visually)

A better way will be to program the computer to be able to “learn”: So you “feed” it 1000+ pictures of cats, and a 1000+ pictures of dogs, and let him come up with he’s own rule-set. The “learning” part here is very precise and limited to ONLY THIS TASK. This same model will not be able to go and play a game of chess. These types of learned task computers already for quite some time are better now to identify objects in pictures than are humans (in terms of accuracy)

I think most layman are not aware at how fast or powerful some of the new ML techniques have become. Although stuff like “general artificial intelligence”(good at many tasks) is still far off.

Excellent explanation. And Lulu and mj can find a really good working example of this in Google Photos.

Give it a few example photos of your kids, and it will instantly trawl through your entire collection and bring just these up.

It’s amazing the sheer variety of “difficult” photos it correctly identifies.

And both Lulu and mj would profit from paying a visit to YouTube and listening to the latest Google research demo of an AI robot caller placing an UNSCRIPTED call to a completely unsuspecting restaurant, and convincingly making a SUCCESSFUL booking with HUMAN, who thinks they’re taking a call from another human.

This stuff is already being trialled at call centres where you know, the HUMAN experience reading off a limited script is often .. er… NOT so intelligent.

End of comments.



Subscribe to our mailing list

* indicates required
Moneyweb newsletters

Instrument Details  

You do not have any portfolios, please create one here.
You do not have an alert portfolio, please create one here.

Follow us:

Search Articles:
Click a Company: