In this post, we discuss a prominent topic: How artificial intelligence (AI), and especially personalization can better serve customer experience – either your customers’ experiences or your experience as a customer.
As a Data Science Architect at Aliz and a former Senior Data Scientist in the Market Research industry, I’ve handled many companies’ Recommendation Engines from different verticals. I’ve worked mostly with retailers, e-commerce marketplaces, including FMCG brands or airlines.
When companies mandate Aliz to enhance their offers with AI, their motivation is not altruistic. Their key performance indicator (KPI) is an increased customer conversion rate at least, a sales lift at best; even if customer satisfaction is nice-to-have, it remains secondary.
In line with this observation, we discuss how AI can improve customer experience with the genuine behind-the-scenes motivation of selling more.
From the AI Index Report 2018, you can see that since 2015, companies in the IT sector have been increasingly investing in AI and Machine Learning at the cost of Cloud and Big Data.
These tendencies are similar for non-IT companies. We’ve been facing the rise of AI and Machine Learning since 2015, even if investments in Big Data were still prominent in 2017 while Cloud was losing some interest at a slower pace than in the IT sector.
All these tendencies are quite legitimate since Cloud, Big Data, AI, and Machine Learning move within the same logical flow.
First, we would not have been able to collect massive amounts of data – Big Data – without Cloud capabilities. Generally speaking AI (and more specifically Machine Learning) could not have improved (i.e., in terms of model performances) without tons of data to explore. There is no denying that data is nowadays considered a pure commodity worth more than oil.
Now let’s take a look at AI from the end-users’ side. We, as customers/consumers, are served by technology. What would our lives be like without technology? Without IoT? Without the Internet? Technology is embedded in our day-to-day lives, our smartphones, the social media content we scroll every day. This technology is based on one paradigm: the Web.
Back in 1992, Al Gore, US VP in the Clinton administration, pushed legislation that funded the expansion of ARPANET into the Internet – a Web 1.0 that was static until the 2000s. Two opposing sides were differentiated at that time: publishers creating content vs. consumers consuming content.
In 2004, the terminology of Web 2.0 was democratized by Tim O’Reilly to refer to a more dynamic Web where the frontier between publishers and consumers became blurred: the same people could now create content and consume it. We were on the verge of social media.
Where are we today? There is consensus that we entered the era of Web 3.0 in 2015. The technology we are currently using follows this new paradigm.
Despite some technical differences from the former Webs, Web 3.0 is largely the product of AI, an AI serving us as end-users – with smart apps, a semantic web, behavioral advertisements, or behavior and engagement tracking for greater purposes.
AI has become a new standard. It is embedded in our lives through the technology we use: Google, Gmail, Amazon, eBay, Facebook, Spotify, YouTube, Instagram, Pinterest, etc. each and every medium we consume content from.
AI is almost everywhere. It is even physically in our homes if we have devices like Alexa or Google Home.
One prominent aspect of how AI can improve customer experience is recommendations. Recommendations are the logical continuation of the search act.
Ten years ago, we spent a lot of time typing and retyping in search bars. If search bars are still there and we are still using them, we also rely on recommendations to serve up massive amounts of data for us to consume and use search bars less frequently.
Imagine if we had to search for all the relevant content displayed in our Facebook News Feed, Spotify recommended playlists, Netflix browsing page, YouTube home page, or Instagram Explore tab.
We, as customers, are usually keen on new suggestions to broaden our experience. This is especially true for fast-moving goods (e.g. FMCGs or streamed content) since they are cheap and quick to consume. We are usually ready to experiment with new content since it costs us nothing except time.
However, for more durable goods (e.g. electronics, clothes, travels), customers are not so open to trying out suggestions: these goods are not intended to be replaced in the short term and they are more expensive. For these products or services, recommendations should be more targeted; additional insights are usually leveraged to convert people into purchasers.
In Part II of this series, we will dive into how to start out with a recommendation engine: what to consider, how to start building it, and what factors to weigh up when thinking about personalization from a Machine Learning perspective. Stay tuned, and subscribe to our newsletter to make sure you don’t miss it!
