At this year’s Amplify conference in Las Vegas, there was one major theme echoing throughout the MGM Grand — cognitive is a journey, not a moment. In my role as a technology strategist in IBM’s Cognitive Incubation Lab for Watson Customer Engagement, the idea of the cognitive journey rings very true but my view comes from a slightly different perspective, a technology one.
Cognitive Incubation Lab
If you’re not familiar with our cognitive incubation lab, it’s the place where IBMers go when they have a cognitive concept they want to morph into a solution capable of driving real business transformation. Our team has the honor of looking through some terrific ideas. Of course with that comes the struggle to decide which we should develop to help advance the cognitive journey.
Aligning Search Results with Intent
The latest example of our work debuted this week at Amplify. Tentatively called Search Insights, it uses semantic intent to understand buyer modes derived from a user’s natural language query. For example, a shopper might be researching, discovering new products, looking for help, or ready to buy a specific product. It then ranks the search results to align with the intent. As more and more users interact with a retailer’s site-level search, the system learns from their behaviors to reorder the results appropriately for future similar queries. Search is a great use case for cognitive.
How Does Cognitive Make Recommendations?
The same goes for recommendations. A great synergy exists between cognitive and recommendations. Why is that? It’s because cognitive is like a never-ending feedback loop that sifts through vast amounts of data, from structured to unstructured or dark data that includes everything from a tweet to an entry in an offline purchase transaction log. Through understanding user’s behaviors and gathering insights, evidence-based recommendations can be uncovered which can be used for almost anything, from delighting consumers with new products and services they might never have discovered on their own or helping a business identify potential improvements to key processes.
For example, a cognitive-powered solution could examine past sales trends, inventory levels, and developing weather patterns to recommend a retailer’s marketing team run a promotion on winter clothing items to regions that are experiencing extended winter conditions. But that’s not all. The solution could also recommend pricing and offers for the item — a price point that will entice customers to buy while adding to the retailer’s bottom line — while providing guidance on which stores should increase inventory and which supplier it should be sent from.
Cognitive Draws from Many Sources
And, since we focus on the cognitive technology journey, we already see opportunities to expand on this type of capability and provide even more personalized levels of recommendations by, for instance, tapping into a person’s social network.
Imagine your online looking to buy a toy gift for your son. You ask a cognitive powered service for toy recommendations for a 7-year old boy who likes Ninjago and Star Wars. The service comes back with several Lego sets based on your request as well as past purchases you’ve made. But that’s not all, by tapping into your social network, it finds friends who’ve bought these items as well as their reviews. Now thanks to cognitive, you’re not only getting ideas relevant to your needs but feedback from those very close to you, the people you trust most.
Cognitive – the only true personalized recommendations
And like any good team, cognitive is always working, modifying its recommendations based on the latest data and user behaviors. By working from the latest information, cognitive is always able to provide recommendations that are relevant to the moment.
Now, of course, cognitive isn’t the only solution that can help in these areas. In fact, I have people point out that there are other APIs available for developers which provide recommendations. That is true. But so is this — cognitive technologies are the first that are truly personalized, and, when it comes to recommendations that is a real game changer.
Share this post: