AI will impact the way brands interact with consumers online...

How? Over to our guest author.

It's easy to think of AI as robots waiting to steal our jobs. Personally, in just about every debate, I have always propagated men over machines because men come with heart and machines come with processes. But the reality is so much more nuanced than that. Over the past few years, I have come to understand that advancements in AI will bring about a rather exciting change in the professional landscape and as one of the biggest business trends of our times, its impact on every aspect of business will be immense. Yet the one that interests me the most as a marketer, is how will it impact the way brands communicate with consumers?

Iyer Premkumar

Let's take a look at how AI, coupled with many of its subsets, will not just help brands deliver content better, but manage contextual intent and, most importantly, bring in better efficiencies in the way they operate.

Let's begin with the DeepText program.

DeepText is Facebook's attempt at understanding the many ways in which text is used. While decoding the intent behind this text, Facebook intends to use it to deliver personalised content/ ads. Imagine you're chatting with someone on Facebook messenger and you mention that you will be leaving in a bit or will book a cab soon; Facebook, using DeepText, can recommend a cab ride service for you. With DeepText and its other AI projects, Facebook is trying to make huge improvements in how content is shown to users. The objective is simple. If every post is interesting to the audience, they will spend more time on the platform, making them a captive audience for businesses advertising on the platform.

Exciting isn't it? Predicting consumer behaviour and predicting it correctly, has always been a pursuit for brands. Now, while there is no magic potion to this, with AI, brands seem to have found a credible and sustainable way to develop a far more cohesive and coherent understanding than ever before. And AI coupled with this understanding of consumer behaviour can be seen to have a visible impact in the way brands are conversing with consumers.

At a functional level, AI today is making listening tools more accurate and effective. Humans being humans and with written text making it difficult to ascertain tone and subtext, it gets extremely difficult to contextually understand and read beyond what is written, for example - sarcasm. Add to that, communication and language (especially in India) is not standard and each and every consumer communicates with the brand the way they deem fit or feel comfortable with.

This is where Natural Language Processing (NLP), a subset under AI where maximum value can be brought into social listening, comes in. Today, 70 per cent of the conversations are being tagged as 'Neutral', which renders them meaningless to marketers. However, if we carefully read each and every mention, they can all be classified as 'Positive' or 'Negative'. This is where NLP can work wonders. When deployed correctly with machine learning, it can contextually evaluate voice and text beyond face value. The combination of NLP and machine learning can ensure sentiment classification with solid accuracy.

So what is Machine Learning? It is another subset under AI dealing with the algorithm that forms the core for these tools. Machine Learning processes complex data and simplifies it for end users using set parameters to predict usability for us. With machine learning, we can determine patterns for categorisation. By sensing human inputs and usage patterns the tool trains itself to see patterns in imagery, text and videos, allowing for better indexing and categorisation.

The idea with machine learning is to improve performance as humans start using the system more. The best sentiment algorithms can be 80 per cent accurate, but that number can go up to 90 - 95 per cent when machine learning is applied.

The other benefit AI offers is that it allows scaling of digital listening operations with increased speed. Today, conversations are increasing across platforms daily and beyond a point, it will be almost impossible to scale quickly to match sudden increases in conversations. Or if your business needs a 24/7 setup but due to budgetary reasons it is not possible, intelligent chat apps can be put on steroids using AI to make sure customers aren't unattended. For example, Facebook has a bot for digital customer service and enquiries. Microsoft has a bot API for Skype. These chat apps can take care of the consumer's needs to a great extent when your shutters are down.

We are also seeing a change in the way customers prefer interacting with brands. It is almost undeniable that most Millennials prefer non-human interaction support. Being tech-savvy, they prefer solving issues by turning to FAQs, online chats etc. With the younger generations expected to be more independent, the ability to scale up for interactions using intelligent AI will be crucial and game-changing. With the ability to decode emotions behind voice, read text beyond face value and understand images to create a personalised and more human-like approach, apps will be able to converse far more realistically than ever before.

So we find ourselves back where we began - will or rather, can AI engager or replace Humans?

Definitely not!

AI is built to help humans improve efficiencies; e.g. we cannot stay up 24 by 7 managing queries like AI can. We may forget instances, but AI won't. And the fact is un-debatable that AI is still in the fine-tuning stage with a glorious future ahead. Yet, AI should not be seen as a threat; instead, as marketers, we must embrace it and use it to complement our weakness to deliver the great experience we promise our consumers.

Integrating AI into our business is not like switching the AC on. It's a process, a journey that needs commitment. And now is the perfect time to begin because, like every great partnership, it requires seamless integration (of technology and data) to not just function, but thrive.

Remember AI is not a magic wand and the "GiGo" concept of computing, where the quality of output is determined by the quality of the input, applies to AI perfectly. Whether we agree or not, AI signals a massive leap towards hyper personalised marketing and it's time we took that first step.

(The author is director, digital command centre, Gozoop, an independent digital agency).