The what, how and why of AI

Data is often the most under-used asset a company has, and a ‘data-first’ approach forms the foundation of AI, says Shakeel Jhazbhay, GM: Digital Business Solutions at Datacentrix.

Durban, 25 Sep 2019
Read time 4min 10sec
Shakeel Jhazbhay, general manager: Digital Business Solutions at Datacentrix
Shakeel Jhazbhay, general manager: Digital Business Solutions at Datacentrix

Data plays a pivotal role in any artificial intelligence (AI) implementation. In fact, data is often the most under-used asset a company has, and a ‘data-first’ approach forms the very foundation of AI.

This is according to Shakeel Jhazbhay, general manager: Digital Business Solutions at Datacentrix, a high performing and secure ICT solutions provider, who explains that a Narrative Science survey confirms his claim.  

“The survey results state that 61% of companies with an innovation strategy say they are using AI to identify opportunities in data that would otherwise be missed, while only 22% of those without such a strategy could make this claim,” he says.

“Clearly, then, it is critical that local companies have the right data strategies in place – as well as a clearly defined implementation and adoption plan – in order to ensure a successful AI journey.”

What AI is… and what it is not

Jhazbhay explains that AI can be defined as the ability of machines to mimic human-like intelligence; to learn, adapt, reason, problem solve, self-correct among many other actions. “What AI is not is a living entity with real perception of emotions, sensory awareness and high cognitive intelligence that exists on its own, without any form of human intervention.

“There’s no doubt that AI has a critical job to perform within today’s society, and there is a growing awareness from consumers as to where it is used on an everyday basis,” he adds. “We’re seeing the proliferation of chatbots, for instance, as well as a growing use of image, video and speech recognition.”

Gartner has predicted that by as soon as next year the average person will speak more frequently to a bot than to their spouse. In keeping with this sentiment, Salesforce research from earlier this year states that the projected growth rate of the use of chatbots within service organisations is targeted at 136%.

From a consumer point of view, a study by Creative Strategies uncovered that 97% of mobile users are already using AI-powered voice assistants, with 51% of mobile phone owners using voice assistants while driving specifically. Also marketing company, WordStream, says 40% of adults use voice search functions at least once every day.

“AI has become commonplace within our everyday lives, transforming the way we do many things,” Jhazbhay says.

Why go AI?

Businesses that successfully leverage AI in solution delivery to customers stand to gain a reduction in operational costs, increases in productivity, operational efficiencies and revenue, and improved customer experience. AI also eliminates human error, as computers are better suited to handling the repetitive tasks that humans don’t like to perform.

Applicable across any number of market verticals, common business cases for AI include:

  • Banking and finance: AI has been used to detect and prevent fraud, client retention, wealth management, and provide personalised recommendations;
  • Insurance: AI has been leveraged to generate new sales leads, risk analysis for new applications, behavioural policy pricing, claim validation and settlement; and
  • Agriculture: AI has been used to help farmers predict crop yields, which is vital in water restricted areas like South Africa.

Jhazbhay recommends that organisations looking to implement an AI solution should ensure proper planning is done beforehand, taking the following practical steps:

1. Determine the quality of an AI product by looking at past success rates, analysing previous case studies of the AI application in question;

2. Once your business has decided on which route to take, a clearly defined use case must be put together;

3. Then, data sources and availability must be verified;

4. Basic data exploration and analysis must be completed;

5. A model building and validation methodology needs to be defined;

6. The model must be built and deployed; and

7. It must then be continuously updated.

AI plays an important role in allowing businesses today to get to grips with, and derive meaning from, an ever growing volume of data from various sources, Jhazbhay explains. “Organisations are using AI to get the most out of this information through deeper analysis, discovery and progressive learning – a task easily overwhelming to humans,” he says.

“However, AI is intended to enhance human capabilities and contributions by performing such mundane tasks in an efficient, reliable and predictable way, thereby freeing the humans to focus on more complex tasks.

“The good news is that AI should also create more jobs than are displaced, with the World Economic Forum (WEF) forecasting that AI could create 58 million new jobs within the next few years, with a large shift in job quality, location and permanency expected,” he concludes.

Editorial contacts
Nicola Read (083) 269 2227 datacentrix@pr.co.za