Twenty-five years ago this week, Amazon looked less like the future of retail than a casualty of the dotcom boom, when its stock plummeted by 95%. Spoiler alert: it’s recovered a bit since then to become America’s favorite retailer, but it’s a handy reminder from history that predicting losers can be as hard as predicting winners.
Scroll down to read about undecided voters, the increasing use of AI in decision making, and that Massachusetts lesbian bar.
Stats to power your week
🫰Reaching the top of the App Store just after its release, Meta’s Muse is making people excited about the possibilities of agents and agentic commerce. But on the consumer side, agents might be deployed for more mundane tasks. The main use for AI tools when shopping right now is to compare prices (51%). In the same way many currently view ChatGPT as another search engine, for many people Muse might be another price comparison tool. GWI Pulse
🏖️With that said, for what types of purchases might agents be used the most? Even since the pre-agent days, travel planning has consistently emerged in our research as the most in-demand use-case, and our latest survey is no exception with over a third (37%) using AI for that purpose. If nothing else, human travel agents will have an interesting challenge to remind people of their value. GWI Pulse
📈Elsewhere in travel, an intriguing trend has seen Booking.com close in on Expedia’s lead as the OTA of choice in the US. 18% would use Booking.com if they were booking a trip tomorrow, compared to 20% for Expedia. GWI USA
😷Thanks to some delicious reporting from The Boston Globe, an updated masking policy at a Massachusetts lesbian bar has become a viral sensation (even Hollywood has come calling). Masking has lingered in some progressive spaces, particularly in the US, but what actually predicts if someone is worried about infectious diseases? Perhaps unsurprisingly, 56% of Americans that use Bluesky multiple times a day rank disease as one of their main concerns. GWI USA
Meet the audience: Undecided voters
Without GWI data
With GWI data
This week we wanted to profile an audience that will become the subject of much interest as the midterm elections approach: undecided voters. We should confess, AI captured a pretty good idea of our target audience on its first attempt. In fairness, there’s a lot of data around voting intention, so its depiction of a classic undecided voter - a rural mom who spends most of her time thinking about family and not politics - is pretty accurate.
So this week is more of a test-case in showing how an extra layer of data can help understand the different segments of an audience, not just the headline one. From that we meet the others likely to be sitting on the fence come November: young Latina women working part-time in the Midwest, fortysomething tradesmen that don’t trust anyone in power, and open-minded Gen Zers that haven’t found their tribe yet.
Chart of the week
Last week we looked at the changing landscape of what AI tools are used in organizations. This week we’re asking: how are they being used, and how has that changed?
First things first: time efficiency has always been the biggest identified benefit, and continues to be. But movements further down the leaderboard are very revealing.
A lot of the narrative around AI assumes businesses will use it to save costs, particularly on labor, but this has become less important over time (as anyone who’s hit their token limit can verify). What’s striking is how much AI is getting plugged more deeply into systems, becoming an additive factor. So it’s increasingly cited as a productivity driver, and something that actually improves decision making.
We’re still learning how to work with these tools, and it’s clear that effective decision-making goes beyond just asking ChatGPT or Claude what they think, but if businesses are increasingly citing it as a benefit, the extra intelligence appears to be making a real difference.
More from GWI
Everyone’s talking about synthetic data, but how accurate is it really, and how do you do it right? Our Chief Insights Officer Jason Mander will reveal all in a webinar on October 15th - tune in here.
And in case you missed our last webinar on agentic ad selling, you can watch it on-demand here.
