Human collaboration key for business AI adoption

2018-03-11 12:21 - Sophia Liu
Post a comment 0


Johannesburg - As the world launches into the era of industry 4.0 or the fourth industrial revolution, many businesses are looking to various technological innovations and advancements to remain competitive.

One of these business transformations include the increased adoption of artificial intelligence (AI) in marketing applications.

The hype around AI has been massive, as it brings promises of reduced labour costs, optimised production and operations and more efficient timing and delivery, through capabilities such as predictive analysis, propensity modelling and lead scoring.

Whilst AI will undoubtedly bring worthwhile opportunities for business growth, there are challenges to consider before adopting AI for your company.

Firstly, it must be understood that AI platforms for business marketing, usually in the form of machine learning, deep learning or natural language processing, are still in the primitive stages of maturity.

There are plenty of pitfalls which require enhancing and refinement, before it can claim the title of ‘true AI’.

One example of this is programmatic advertising. With real-time bidding and automation in open ad exchanges, it took the digital world by storm and was quickly embraced by many businesses.

Unfortunately, it was soon discovered to be fraught with many loopholes, leading to ad inventory issues, such as ad fraud, bot and domain spoofing and brand safety challenges. 

Human Element

Consequently, it would be wise for businesses to consider a ‘human-centric’ approach to AI. Understanding the current limitations of AI is crucial for business leaders to leverage it for their companies’ needs.

While advancements in AI is happening quickly, the human element is still required to drive Return on Investment (ROI) manually and when AI is adopted, it should be used collaboratively with humans to enhance decision-making, not replace it. Thus, it would bode well for companies to foster an internal interest in AI.

A few examples to achieve this can include sharing articles on the subject, facilitating workshops with AI experts or arranging staff to attend specialised AI conferences.

Last but not least, there is the major contemplation of AI bias. Machines still require people to input data and this data can unwittingly incorporate the human biases of its creators or users.

In turn, this will have a significant effect on its output. If your team do not have the necessary skills to make AI worthwhile and feed it skewed data, then the results will be skewed.

Think of Facebook’s algorithm which decides to show us which posts we most likely want to see. But there will be many posts which we don’t get to see, simply because this algorithm decides against it. 

One way to control this is to implement processes and procedures for data governance.

Another, perhaps more radical idea, is to embark on social initiatives within the company to create more empathy and cohesion, especially if such biases are, for instance, gender, culturally or racially-based.

Having a comprehensive understanding of the benefits that one can reap from this technology is vital to any business’ AI strategy.

Just because certain aspects of it may not work for one company, does not mean that it won’t work for your business.

The secret sauce is in the collaboration with your team to gain an understanding of the ‘how’.

*Sophia Liu is a Johannesburg-based brand communications specialist and media strategist.

* Sign up to Fin24's top news in your inbox: SUBSCRIBE TO FIN24 NEWSLETTER

Read more about: ai  |  artificial intelligence