This post was created in partnership with Fetch
Fetch, a rewards-driven app where consumers earn points for shopping, has teamed up with OpenAI to launch FAST (Fetch AI Semantic Technology), a platform for brands and marketers that leverages its trove of consumer data to provide insight into spending habits.
Using natural-language AI, brands interact with FAST much like OpenAI’s well-known ChatGPT and can ask questions like, “Show me sales revenue in the last three months.” FAST will then compile its data across Fetch’s roughly $200 billion in annual sales receipts and return an easy-to-understand response like, “Sales were $350 million in the last three months.” Marketers can ask follow-up questions like, “How do we compare to other retailers?” to get a more robust picture of sales performance and get real-time insights into spending habits. FAST can even suggest actionable marketing steps with a prompt like, “Give me five ideas for a Fetch points offer to be more competitive.”
The ‘Aha’ moment
Founder and CEO Wes Schroll knew he was sitting on a data goldmine when he started the company back in 2017. “Our average user submits 31 receipts per month. And across 12.5 million active monthly users, that means we’re capturing almost $200 billion a year of sales. Everyone always would say how valuable the data must be,” Schroll said. But data is inherently messy and organizing it back then was tough. It wasn’t until the development of large language models that Schroll had a breakthrough. “As we started to look at some of the evolving AI technologies, one of the things that they’re really good at is sitting on top of large sets of data and making them very approachable.”
For Schroll, the key was making the data accessible for anyone, and the use of natural-language AI was a gamechanger, especially when it came to activating a brand or retailer on Fetch. “Typically, with those [data] questions we’d have to go back to our analytics team, run custom queries, then come back to them weeks later and say, ‘Here’s what we can do,’ versus being able to go into it in real-time. It just condenses decision-making,” he said.
According to Schroll, one of Fetch’s taglines is that “rewards are a better ad” and despite FAST being a tool for brands and retailers, it helps Fetch remain a consumer-driven platform. “The reason Fetch users open the app every single day is they’re looking for ways to earn points, which are advertisements from our partners,” Schroll said. “The better we get at targeting, the more relevant the consumer experience is. Because when you open the app, you’re always seeing something that’s interesting to you and gives you a chance to earn points.” In doing so, brands and retailers can target ads to their specific market and drive measurable outcomes, while consumers get a “magical experience that’s very relevant to [them].”
Getting the data right
As powerful as FAST is, accuracy still remains a huge question when it comes to implementing AI technology and it’s why Fetch still has a human touch. “Our quality process for FAST combines the excellence of our human data experts with automation to deliver consistent results aligned to our high bar for our data,” Schroll said. “Thousands of prompts have been audited by humans to set the standard for accuracy and ensure quality. Humans remain in-the loop today, combined with automated review systems that are measured against our high-quality standards.”