Apple’s release of Siri and Google’s integration of semantic intelligence into search results signalled a shift in our online user experiences. The web had started a quiet evolution into being a Semantic Web. While that term has lost its lustre, we’re now all familiar with the same concepts rechristened as Big Data and AI (artificial intelligence).
It’s a shift that’s constantly happening in the background of each of our daily online experiences. Every eCommerce transaction we complete becomes another signal – more data to fuel the fire.
This isn’t the first or even second time that AI is heralded as the next big thing, but this time it looks like the computing power is finally here to bring it into multiple consumer-grade products (think Alexa, speech to text,
Big data is changing eCommerce
Advances in technology are powerful instigators for social adaptations & cultural evolution. The phenomenon that came about thanks to the shift from the original publishers’ web to the user-generated web 2.0 such as blogging, the social web, and crowdsourcing are all evidence.
These social shifts created reams of data that yielded a new technological opportunity big data sets large enough to allow for machine learning at scale.
Coupled to increasingly more powerful processors and a growing abundance in talented engineers a new generation of AI products is being released to the market
How is our eCommerce experience shaped by AI?
For the most part, our purchases on today’s web are largely modelled on offline commerce & fall into two main categories:
- Impulse buys
- Planned purchases
Because impulse buys are by their very nature – unplanned, I’d like to focus on planned purchases. For this post, we’ll use an example of anyone who’s had a baby is familiar with stroller hunting.
How is AI eCommerce different?
To understand how AI-powered eCommerce differs from our current experiences, let’s review what we’re familiar with today.
Hunting for the perfect baby stroller, the one best suited to our needs usually involves some or all of the following activities:
- Consulting with friends
- Discussing the purchase on forums & social networking sites
- Googling to get a grasp of the stroller market
- Comparing stroller prices with price comparison sites
- Hunting for bargains
- Posting “Wanted” ads on second hand & free-swap sites
Interestingly all these activities share the following attributes:
- They’re motivated by a clearly defined & obvious need
- They require the usage of specific web services
- Repetitions of the same tasks are often required if we are to get the most value from our activities
- A fair amount of time is wasted
- Aggregate & analysis of the data requires man-hours
The bottom line is that today’s Web saves us most of the effort of leaving our homes for purchasing research, but online shopping experiences are still not much more than digitally enhanced bargain hunts, & when we factor in the time spent finding a “Bargain”, our purchases often cease to be anything that can be classified as one…
The good news is that AI is already changing most of that.
Ecommerce stroller shopping – AI style
An ideal AI is one that sees the web for what it is – Data that is imperfectly defined, structured & linked, but is accessible.
With sufficient time and computing power to drive machine learning, this web of data is parsed to make product research, price comparisons & bidding all completely automated. Here’s what shopping for a stroller looks like to an AI agent.
Consumer intent – explicit or implicit = big data
Pregnancy has inevitable cues in our own online activity – Tagged pregnancy pictures are uploaded. Gripes about morning sickness start appearing in online statuses. These are all explicit cues.
But there are implicit cues that give away just as much information, if not more – Visits to parenting sites leave cookies on browsers, as does browsing for baby clothes. Shopping patterns change as certain products are no longer purchased, while others start popping up in our cart with regularity
As the pregnancy progresses so does the frequency of our related online activities. This creates a pattern, which is replicated across millions of pregnancies across the planet.
AI parses the cues
Each one of our actions by itself is nearly inconsequential, but to an all-aggregating & all-reasoning AI the cumulative effect of all of them means only one thing: Someone is having a baby …&, therefore, they’re ripe for pregnancy-related content & advertising.
AI optimized eCommerce
An AI doesn’t waste our attention by advertising products it knows (from our previous purchases & bank transcripts) we can’t afford. That’s inefficient. An AI saves consumers the effort and frustration of sifting through irrelevant offers by targeting every consumer with ads that take into account their particular unique needs & circumstances. Each person is offered only stuff likely to be within their price range, & supplied by the vendors that are closest to them.
AI optimized advertising
Much of the advertising we’re served is already AI optimized – Ads for items we’ve already purchased are replaced with ads for items that compliment & augment them.
We’re presented with ads for vacation dates we’ve entered on Agoda or Booking.com, but once those dates have passed, we’re no longer exposed to those ads.
Where will AI-powered eCommerce take us next?
The big difference between eCommerce today & AI-powered eCommerce is that AI will save us the need to deal with all the mundane tasks we have to deal with ourselves today.
We’ve already uploaded enough data about ourselves to Facebook, LinkedIn, Amazon, etc. to create a very accurate picture of who we are & what we’re likely to want and buy. Why not enjoy having products and activities optimized for the US simply appear magically on our phones? Why, instead of hunting for bargains, have AI packaged bargains hunt for us?
We’re not fully considering the implications of this shift at the moment. All this wizardry has a cost. For one it means that more power than ever will be consolidated in the hands of those companies with the financial clout to develop the technology required to be on the cutting edge of this shift. The competition to create ever more powerful AI is on, and it’s happening at the state level
The 2.0 evolution of the web was primarily democratization of the platform. What we’re experiencing now is quite the opposite – The biggest fish in the pond are leveraging big data and computing power to create technologies that give them an even bigger advantage over their competition. It’s a virtuous cycle that keeps them ahead of an ever-widening gap from the competition, to the point that competition becomes impossible. We cannot have another Google or another Amazon – it’s simply impossible to do.
The survival of smaller businesses in this new environment requires entirely new thinking – What that looks like is hard to tell, but necessity is always the mother of invention – I for one take hope from that.
