There is huge excitement about artificial intelligence (AI), but most organisations are still battling to understand what it really is and how it can add value to their business operations.

By Kathy Gibson

“AI has a lot of promise and it is going to be everywhere,” Magnus Revang, research director at Gartner, told delegates to this month’s Gartner Symposium in Cape Town.

However, it will take decades before it becomes pervasive, he adds.

“It is so easy for us to jump ahead of ourselves. We think things will happen faster than they will.”

The people who have done nothing about AI so far are nervous about it.

Most of them (79%) cite fear of the unknown; 63% are battling to find a starting point; 48% are waiting for a vendor strategy and 40% say they’ll start when AI reaches enterprise maturity.

Revang urges organisations to have patience when it comes to AI investments.

CIOs have high expectations from AI. In fact, it was ranked highest out of five business areas – higher that Internet of Things, 3D printing, cloud-based API platform and blockchain.

Revang warns against organisations taking panic decisions about AI, and deploying systems that are almost certainly doomed to failure.

“The fear of missing out will lead to some extremely poor decisions,” he says. “Companies that rush will implement poor decisions.”

A lack of awareness will lead to overblown expectations and disappointments, Revang adds.

“The word intelligence is so misleading – because the only intelligence we know is human. But machines are nothing compared to this. Our reference frame leads us to borrow human qualities to understand technology.

“But it’s better to think about amazing innovations. Don’t add human capabilities into it and get misled.”

Some of the myths he believes need to be killed include:

* Buy “an AI” and you’re ready for anything;

* It’s easy to train DNNs and NLP;

* AI learns on its own;

* AI sees like people do; and

* You only need one AI platform.

“If you haven’t started with AI, you are not alone,” Revang adds.

So far, 14% of CIOs say they have no interest; 35% have it on the radar, but haven’t planned any action; 25% are in medium- or long-term planning; 21% are in short-term planning or actively experimenting; and 4% have already invested and deployed.

“But with so many experimenting and planning, we need to start doing something in order to keep pace.

“And, if we do the right thing, we will be able to scale it and get business value.”

Organisations need to find the areas that will demonstrate the value of AI, Revang points out.

Gartner believes there are six ways AI will enter the enterprise:

* AI platform as a service;

* Niche solution, in many narrow services;

* Custom projects;

* AI-improved applications and suites;

* AI-enhanced customer-facing channels; and

* Employee consumerisation.

Gartner predicts that, by 2021, AI augmentation will generate $2,9-billion in additional productivity.

“The best case for business value will be in focusing on the customer experience,” Revang says.

This means all customer touch points, though, not just the call centre. AI will be used in all aspects of customer growth and retention.

The other business value driver is cost reduction. AI could be employed to boost process efficiencies, improve decision-making and automate more tasks.

By 2021, however, the most value from AI will be in driving new revenue.

Revang adds that the business value of AI will be proportional to how thoroughly an organisation re-invents its business.

“At some point you will be asked to do that. But you need to have a lot of experience, which is why you need to start working now on the early AI deployments,” he says.

Early deployments will help organisations to learn about what is important to them.

“You can also look at others and see if you can adapt what they have done,” Revang says. “And don’t look at just your own industry.”

It’s vital to know what data the organisation has, and what it could tell them.

“With all this you get the possibility that can be assessed for feasibility, or provided with data science.

“And then you will have something you can use.”

The next step, says Revang, is to listen to consumers, and understand their preferences on how to interact with AI.

For instance, most users don’t want to be told what to do, but are looking for help or assistance. “They want the ability to be the smarter one in the relationship between human and machine.”

It’s important to pursue the human-machine symbiosis to maximise value, Revang adds.

One of the most important business case considerations, however, is the data.

“This is the most important strategic asset you have in the age of AI,” Revang says. “Everything else is tactical.”

This means organisations need to spend more time on data, training and algorithms.

And to do this, they must account for the importance of data; select business problems that have enough supporting data; and budget for adequate data preparation efforts.

Revang concludes that organisations need to be patient. “But don’t wait. There are several generations of AI-enriched applications to come. Focus on the outcomes.”

He recommends that organisations aim to follow quickly as AI applications are deployed, and learn from all industries.

“And understand the options: you don’t have to build things yourself. A lot of these applications can be bought off-the-shelf.”

It’s important to interrogate claims, Revang adds, and the ability to do this will follow from fostering AI literacy within the organisation.

Share This