AI and ML are Taking Digital Marketing to the Next Level
Updated: Feb 21, 2019
In an presentation by Denyse Drummond-Dunn, she shares her ideas and takeaways she researched and discovered regarding AI and ML in the digital marketing sector. Read further to find out more in depth of the presentation. Article by- customerthink.com
"I presented last week at an exciting, forward-looking conference in Fort Lauderdale, USA. It was ITEXPO, a successful conference celebrating its twentieth anniversary this year! The huge turnout of thousands clearly shows the value that both attendees and exhibitors get from it.
This year the programme included a new stream, the Future of Work and that was the one in which I was invited to speak. Before summarising what I presented, I’d like to share some of the ideas and takeaways that I discovered about digital marketing and the impact of AI (artificial intelligence) and ML (machine learning)."
From text to voice:
Most of us have grown up with text communication, but Gen Z, those born after 1996, are more comfortable with voice. They are less formal but far more impatient than previous generations.
They expect Alexa, Siri, Cortana and similar voice-activated personal assistants to be available whenever they have a question. With this type of search expansion into daily life, being on the front page of Google is no longer good enough. You have to be the number one answer to their questions!
AI is not one technology:
Despite what digital marketers may have hoped, AI is not the solution to all our problems. It is simply a series of technologies addressing various current and future customer needs.
Unlike normal analytical processes, using AI needs developers and users to start with the end in sight. Knowing what we are looking for, rather than waiting to see what the analysis brings us, needs a very different thought process. The questions asked become as important as the answers received, if not even more so. Therefore it is advisable to make them the best you can possibly ask. Your digital marketing has everything to gain and nothing to lose.
AI is far from 100% accurate:
AI is still in its infancy, despite great leaps forward in some areas in the past year or so. For example, language translation is still far from accurate today, but that doesn’t mean it’s not useful. Anything that moves us toward increased customer satisfaction from our digital marketing efforts is great. However, we must understand their limitations and not be fixated on perfection.
One of the biggest challenges is siloed data – still! It is easy to see that the more information sources we integrate, the more accurate our platforms are likely to be. But until we finally break down our internal silos AI will not be able to deliver to its full potential.
Taking the robots out of humans:
Robots are not new. Henry Ford was one of the first to realise the advantage of taking robots out of humans. In other words, gatting machines to do the boring, repetitive tasks done until then by people.
Today we need to consider the digital workforce as also an HR challenge and not (just) a technical one. Humans are not upskilling and progressing as fast as robots are. This is the real cause of any work losses that may happen as automation rolls out.
The future of work
Now that I’ve touched on the elephant in the closet, that everyone is secretly scared about, that of job losses, let’s talk about employment. The future is not so much about replacing workers, as in expanding and amplifying their work through the use of AI.
The future will be a world of work plus AI, not work minus AI. When, not if, robots take on many of our current tasks, humans will need to supplement their knowledge with soft skills, ones that AI can’t replicate, at least for now. This is why I, like many others, refer to AI as augmented intelligence rather than artificial intelligence.
One area that will certainly need a tremendous amount of human input is in speech analytics. You probably don’t realise it, unless you’ve learnt another language or two, but speech has an enormous diversity in the ways to say the same thing. Just ask any owner of Alexa, Siri or Cortana! Sometimes their responses are hilarious, at least at first, but these quickly become irritating and frustrating, when you can’t make yourself understood.
If robots are to understand humans, then these alternative expressions need to be programmed in, before being understood. Although machine learning may speed our progress, the foundations must be identified and created by humans.
AI and care centers
Most businesses have customer service departments and many are jumping on the bandwagon of requesting AI. However most don’t really know why they need it! The case for AI has to be put into terms of its business impact and relevance in order to be valued beyond mere “modernisation.” Just ask anyone who has chatted with a bot or gone round in circles on self-service phone lines! So many corporations today have increased their technology but have not improved their customers’ satisfaction.
AI is already proving to be of great value in following and analysing customer service connections. A supervisor can’t listen in or read every exchange, but AI can. However, as previously mentioned, understanding speech is still in its infancy, especially when it comes to sentiment. An agent will quickly sense when something is wrong or an answer is unsatisfactory, even when the customer is saying everything is alright.
The customer journey that led to the connection, is just as important as the call itself. This is where total integration of all touchpoints is vital. The customer already sees them as such, but most companies do not. This leads to irritation when a customer must repeat their details and experiences with each new customer service agent.
It could be so easily eliminated, by simply integrating multiple data sources and then assessing the customer’s “effort” in getting the answers they are looking for. The greater the effort has been, the quicker a solution should be found.
I believe that not taking the customer’s perspective here is the root cause of this less than satisfactory situation today. Once again, adopting a customer first strategy is the answer. If you would like help with this or don’t know how customer centric you are today, why not contact C3Centricity and complete our complementary C3C Evaluator™?
Customers in developed markets already have far more interaction with AI than they probably realise. However, when developing chatbots it is important to allow for far more variation than we are aware of. The challenge is not only understanding the variations in vocabulary mentioned previously, but also colloquialisms, spelling mistakes, acronyms and alternative expressions.
Therefore, instead of aiming for perfection, by brainstorming all possible variants, our time is better spent in identifying the 20% of variations that cover 80% of the cases. Ideally we should first collect information and then analyse what the company is likely to receive most of the time. Perfection is once again the enemy in progressing the use of chatbots.
We also need to be transparent about when chatbots are being used. It may be a good idea to make them respond in a friendly way, but pretending to actually be a human is not a good idea. Customers will eventually understand that they are exchanging with a chatbot when the responses they are getting do not meet their expectations.
AI and taking digital marketing to the next level
After all these intriguing sessions, it was my turn to speak. Luckily I was taking a far more practical approach to digital marketing, AI and ML, which I am happy to say was met with enthusiasm. The audience were fascinated with my hands-on perspective and had loads of questions and comments at the end of my talk.
I would be delighted to share my slides with any reader who would be interested in seeing them, but to summarise my main points:
Digital marketing has made our communications’ media choice even more challenging. There are far more channels than ever, many being used concurrently, especially by the under 35’s (for example TV and the internet).There are more brands vying for space online. The relative cheapness of advertising on the internet means that those that didn’t have access to traditional media because of their high costs can now communicate.Customers are more demanding and expect real-time responses to their questions, and ever shorter delivery times for purchased goods.AI and ML can improve digital marketing through predictive intelligence, content curation / creation, dynamic pricing, and especially by improving the customers’ overall experiences.Digital is best used as an amplifier of traditional media, and when connections need to be more individualised, relevant and timely. This is not always the case, so choose wisely.
It is exciting times for marketing with all the opportunities that technology, AI and ML offer us. However, we are still faced with many of the same challenges we always have been. Essentials such as knowing and understanding our customers more deeply, and removing the siloed information hubs within organisation, remain critical.
Without finding solutions to these, digital marketing will perhaps be cheaper in terms of investment, but could become a more costly exercise and no more effective. What do you think?