AI-powered live selling
Live shopping,
without the limits
of a human host.
Live video is where attention is heading — the same shift that made short-form video dominant is pulling shopping toward real-time, live hosts. Almost no brand can run that today, because it takes a real person, live, on a schedule, indefinitely. Lily is the first AI host built to run that show for you.
- Clearly labeled as AI, always — on screen, every stream
- A custom pitch for every product, not one script for the catalog
- Free pilot stream before this is a paid product
Concept preview — illustrative, not a live screenshot
The problem today
Attention is the actual currency on social platforms right now — and most of it slips away the moment a viewer scrolls past.
Live video is the format that holds attention longest. Running it well means a real person, on camera, on a schedule — which is exactly why most growing brands don't do it. Every unanswered comment and every scroll-past is attention you already paid to earn and didn't convert.
What Lily does today
Not a demo. A real host, running a real stream.
A pitch for every product
Every item in your catalog gets its own sales script, crafted for that product specifically — not one generic pitch recycled across your whole line.
Answers real comments, live
Viewer questions get read and answered out loud on air, so people asking something feel heard instead of scrolling past unanswered.
Tells every viewer where to shop
Every stream points viewers straight to where they can buy — consistently, every single time she's on screen.
Never has an off day
No off-script moments, no missed shifts, no training required before the next stream. She shows up the same way, every time.
Where this is headed
Every stream makes the next one sharper.
We're building toward a future where any brand can have a live selling presence, anywhere, all the time. Lily is the first real product proving that out, starting on Instagram — and she's not static. Which products, which words, which moments actually hold attention: that signal is meant to feed straight back into how she sells for you, so every stream gets sharper than the last.
Why this works
The research points the same direction, again and again.
Directional findings from live-commerce research — used here as
context, not as sourced statistics (full notes, flagged unverified:
gtm_review/research_evidence.md).
Interaction, not video, drives engagement
The engagement lift in live commerce research keeps tracing back to real-time interaction — comments, questions, instant replies — not the broadcast itself.
Brand-run livestreams build stronger connection
Interaction and synchronicity in brand-owned livestreams are associated with meaningfully stronger brand engagement than standard browsing.
Participation tracks with watch time
More comments and likes during a session are associated with longer watch duration and larger audiences — the interaction itself keeps people watching.
Demonstrations build purchase confidence
Live demonstrations reduce perceived purchase risk by letting people ask questions and inspect a product in the moment, not after the fact.
Engagement — not discounts — drives conversion
Social interaction, trust, and entertainment explain much of live commerce's success in the literature — price cuts alone don't carry it.
Major retailers are leaning in
Live shopping is moving from experiment to strategic channel for large retailers expanding programs across new markets.
Cost, for context
Most brands can only staff a few sessions a week. We're not built that way.
Host, producer, comment moderator — the minimum crew for one real live-selling session. Most teams can only run a few of these a week.
We're building toward something meaningfully more accessible than a production team, with the ability to go live far more frequently — not a 24/7 claim, a cadence claim.
Honest about where this is today
Lily is clearly labeled as an AI host, always.
We think that's the right way to build trust, not something to hide. This is early — we're looking for a small number of founding brands to help build this with us, not just trial it, starting with a free pilot stream before this becomes a paid product. Your feedback would directly shape what we build next.