AI / Culture / Cybersecurity

Artificial intelligence comes to Christmas shopping

A Brooklyn/Manhattan duo made a bot that will talk you through your shopping while you're on Facebook.

Artificial intelligence comes to personal shopping. (Image courtesy of Gidi)

A brand new piece of technology built by an East Williamsburg/Upper East Side duo wants to be your holiday personal shopper.
It’s called Gidi and it’s easy as heck to use. From the website, you sign in to Facebook and the bot starts asking you questions in Facebook Messenger about the person you’re buying for. You answer the questions, the bot sorts through its database of 700 gifts, and gives you some recommendations.
Gidi is the work of East Williamsburg-based Jesse Green and Upper East Sider Max Klaben. Klaben had the idea for it a year ago, when he played around with the idea of having a text-based service where gift buyers could connect with personal shoppers for recommendations. More than $600 billion of goods were returned in the weeks after Christmas in 2014, according to the Wall Street Journal, at significant cost to retailers, time spent by people and not an insignificant amount of hurt feelings on the part of gift-givers. A text-based personal shopper was a fine idea, but, because of the human labor necessary, one which would be hard to scale.
In April, Facebook changed the infrastructure used on its Messenger, making it much easier for companies to talk to people. It was then that Klaben, reached out to his friend from high school back in Ohio, and had the idea to turn his personal shopper into an artificially intelligent bot using Facebook Messenger as the platform. Scaling problem: solved.

Figuring out what to get my dad.

Figuring out what to get my dad. (Screenshot)

After graduating from a business program at Baruch, and after Green finished a stint in The Data Incubator, the duo got to work building the program — quizzing friends about themselves and their preferences, and scouring the web for great gift ideas to build up a database.

“The problem with bots is that they don’t work as well if you bite off more than you can chew,” Klaben said in an interview. “That’s why we wanted to make a very specific use case and that’s why we are where we’re at now.”

I tried Gidi to find a Christmas gift for my brother. It asked me who I was shopping for, what the occasion was, how much I’d like to spend and the age of the person in question, then asked me to describe him.

“Tell me a little more about your brother,” the robot asked me. “Passions, interests, hobbies. Traveling, hiking, cooking, reading? List as many as you can think of. The more the better.”

I replied that he likes sports and politics and works in healthcare. That he dresses well and runs and seems ambitious. I didn’t put in that he can be a troll on Twitter and IRL, but I could have.

Gidi listed some options, and then more. A few of them were quite good, and stuff I wouldn’t have thought of, but none reached the point to make me want to actually buy. Then I refreshed once more and got a great suggestion: a VR headset. VR is something I’ve covered plenty as a tech reporter, but have never considered as a gift. But more and more VR programming and games are coming online all the time, and if nothing else, it’s a cool toy to play around with for a few weeks. Neat.

So okay, fun service, but how will it make money? It turns out that data is incorporated into the URLs of the links you use to buy the goods on Amazon or wherever they’re listed. For each purchase, Gidi makes a small commission from the retailer. (The industry term for this kind of sales commission is “affiliate marketing.”)

As with most, if not all, artificially intelligent software, the more people use Gidi and the more feedback it gets, the smarter it will become.

But could this piece of technology extend beyond Christmas gifts?

“This bot is relevant for a lot of other industries,” Klaben said. “How do you find the best television? Is it size? By asking a few questions we could find the product that’s most relevant for you. That’s just an example but you can see how it would be transferable.”

Series: Brooklyn

Join the conversation!

Find news, events, jobs and people who share your interests on Technical.ly's open community Slack

Technically Media