Three real-world cases of AI delivering value in enterprise

Toni Sekinah, research analyst and features editor, DataIQ

Antony Bourne, president of enterprise software company IFS Industries, doesn't believe that AI is a thing in and of itself. Instead, he believes that AI is an umbrella term that encompasses a collection of technologies. Six of those are data processing, tangible robotics, decision support, IoT perception, conversational interaction and process automation. As such, when an organisation says it is deploying AI, it imust be using one of those technologies, in his view.

Three clients of IFS deployed the latter three processes - IoT perception, process automation and conversational interaction - and Bourne illustrated how they did so.

Anticimex, a global pest control company, wanted to become more efficient by using AI technology. IFS helped it by putting sensors on its rat traps. This meant that the pest control engineers would know which traps needed to be emptied and which didn’t.

Engineers were able to change the batteries in rat traps only when needed.

The company then started looking at other types of data that the sensors were returning. Information on the amount of power remaining in the batteries as well as their quality and condition meant that they knew when they needed to replace the power sources and didn’t do so unnecessarily. This saved the engineers time and the company money as vehicles were only on the road when necessary.

Robot hand with finger on keyboardThis was an example of IoT perception - using data intelligently to become more efficient. Bourne said: “From an environmental footprint aspect, because they were being intelligent about which traps they needed to service, their driving time and the environmental impact was reduced.”

Cubic, a company that maintains and services ticket machines, is an example of deploying process automation. Cubic wanted to increase the amount of time its machines were up and running ("uptime") and so IFS helped to introduce a system whereby the machine could self-diagnose if it stopped working.

If the machine can not solve the problem, it sends for an engineer.

It would figure out the fault code and assess whether it could solve the problem itself. If it couldn’t, the machine would then send an alert to an engineer that it needed to be repaired.

Bourne added that, in addition, the machine would know which engineer would be the best person to do the job based on past experience. “This meant that machine availability increased by 20%. It also decreased the IT burden on the engineers so they could go and do more value-added activities,” said Bourne.

"Machine availability increased by 20%."

Finally, without giving the name of the company, Bourne explained that a project involving conversational interaction took place within a consumer product review organisation. The company found that between a half and a quarter of all calls received by agents related the first month free service that it offered to new customers, or to changes of details or stock availability. It wanted to utilise AI to give customers a better experience and to be more efficient.

The company used an IFS language processing module to read and analyse customer messages and intelligently work out what action to take. It redirected calls relating to mundane things like cancellations and changes of address to an automated call agent and diverted calls from VIP customers and customers with more complex problems to a human call agent.

"It reduced the number of emails by 35%."

Robot and touchscreen

The call advisers were even given a desktop application that showed them how to deal with certain objections. Bourne said: “It reduced the number of emails the agents needed to process by 35%. Response time improved but, more importantly, not only were the customers happier, but the employees were happier.”

For decision-makers in organisations that want to invest in and deploy artificial intelligence, Bourne has three concrete pieces of advice. The first is to look at your opportunities and think about what your business is trying to do.

The second is to set specific measurable targets of what you want to achieve with the deployment of AI, and not try to go too fast, too soon.

The final piece of advice is take it easy and to outsource AI projects if the necessary skills do not already lie within the organisation.

Antony Bourne was speaking at Gartner Symposium ITXPO.

Knowledge-based content manager, DataIQ
Toni is the senior features editor responsible for the origination of DataIQ's interviews, articles and blogs.