The news: Michaels, the arts-and-crafts chain, launched a conversational AI shopping assistant called "Ask Mike," built on Google Cloud's Gemini Enterprise, and says it went from concept to production in about six weeks.
The numbers: Michaels says Ask Mike has fueled nearly 75,000 conversations, with more than 60% focused on product discovery across its online assortment. The company separately told Modern Retail that shoppers who use the tool convert at more than double the rate of those using traditional search, and that 27% of interactions end in a product click or an add-to-cart.
All of those figures are self-reported, and none has been independently verified.
Why it was fast: The operator lesson is in the build, not the chatbot.
Google Cloud's retail lead Kapil Dabi told Modern Retail that Michaels hit a six-week turnaround because it "had modernized its technology stack and organized its product catalog in a way that is AI-ready."
The gating factor for a conversational layer is catalog and product-data hygiene, not the AI model. The retailers that move fast are the ones whose catalogs were already clean.
What it changes upstream: A discovery assistant steers demand differently than search. Ask a bot to "throw a birthday party" and it pulls a basket of balloons, streamers, and cake supplies together, rather than returning one SKU at a time.
That bundled, project-based buying reshapes the demand signal planners and inventory teams work from, and it rewards accurate cross-category stock data.
The field: Michaels is one of several retailers deploying Gemini Enterprise this year, alongside Macy's, Ulta Beauty, and Home Depot.
The stack is not universal: Lowe's is running a separate assistant built on OpenAI, a sign retailers are hedging across AI vendors rather than standardizing on one.
