Your AI manager
The only employee you brief in one sentence.
Ask about your store, then ask it to change something. Lumnya answers from live data and acts — with your yes on anything that matters.
01Reads
02Shows you
03Waits for your yes
04Acts and journals
A working preview on a demo store. Pick a brief, or approve one yourself — figures are illustrative.
Ask anything
Answers from your live store, not last week's export.
“How are sales this week?” “Which products are about to stock out?” “Who’s at risk of churning?” Lumnya pulls from a fresh snapshot and replies in clean cards — KPIs, tables, and tappable next steps you can act on, not paragraphs you have to parse.
Then act
From answer to action in the same breath.
If you can do it on a page of the dashboard, you can ask for it instead. The chat isn't a lighter version of the product — it reaches the same fields, through the same guarded handlers.
- Pricing — down to the right variant on multi-variant productschat = page
- Refunds — full or partialchat = page
- Discount-code managementchat = page
- One-off customer emailschat = page
- Store settings — full field parity with the Settings pagechat = page
- Triggers — both the manual rules and the smart enginechat = page
- Returns policy and individual return decisionschat = page
- FAQ answers — and importing drafts from your Shopify policies or a public web pagechat = page
- The storefront widget — appearance, behavior, compliancechat = page
- Support tickets — list, read and replychat = page
- Remarketing — preview a sweep before it sendschat = page
- Report schedules, and running a report on the spotchat = page
Confirm-gated by design
It shows you the change, then waits for your yes.
For consequential actions, Lumnya renders the exact diff — before/after price, refund amount, the email it would send — and stops there. Approve it and it acts, then journals what changed and how it was authorized. Refunds, customer emails and return decisions stay behind that gate permanently, however you set the rest.
Price change — ready to apply
What will change
Action journal0 entries
- Nothing written — it’s waiting for you.
Before you trust it
“What would you do?”
Ask Lumnya to simulate, and it runs the real decision path against your real settings, then tells you which areas it could act in on its own — naming them — which ones would still come to you, and what ’s waiting for a decision. Nothing changes. It’s a dry run, not a demo.
After it acts
One click puts it back.
Reversible actions store their own inverse, so undo replays the same guarded path the original write used rather than editing data behind the product’s back. Irreversible ones say so plainly instead of offering a button that lies.
Every hour, Lumnya checks whether orders have stopped, whether the widget is failing your customers, and whether people asked for a human and got nobody. The comparison is your own normal for that hour of the day, so a quiet shop is never mistaken for a broken one — and one problem is one message, not one an hour. It never guesses a cause it can’t measure, and it changes nothing on its own.
Excel, CSV, PDF or a photo of the page. Lumnya reads the lines, matches the SKUs against your own catalog instead of trusting what it read, and lays every cost change next to what’s there today. A line it can’t match is named, not skipped; a price it can’t read unambiguously it refuses to guess. Your yes applies exactly the figures you were shown — and the whole batch undoes in one click.
“Keep margin above 30%.” “Recover more carts.” Lumnya turns that into a plan you approve once, then tracks it against real numbers and reports what moved. A goal is a standing brief, not a new permission: every step it takes still passes the same caps, the same gates, and the same things it will never do alone.
No orders since 14:00
By 16:00 on a normal day you’d have about nine.
one message · no repeat at 17:00 · cause not guessed · nothing changed
supplier-prices.xlsx
5 lines · read, then matched to your catalog
SC-90 — no such SKU in your store. Named, not skipped.
BN-S — “4.250” reads as 4.25 or 4,250. It asks; it doesn’t guess.
Standing brief
“Keep margin above 30%.”
- Plan approved once
- Re-price four low-margin variantsgated
- Stop stacking codes on full-price itemsgated
- Report what moved, every Mondaygated
Demo store; figures are illustrative.
Real examples
What it looks like in practice.
“Set the Large variant to €30” → preview → you approve → done. “Refund #1043 and apologize” → draft email + refund preview → approve → sent and logged.
Demo store; amounts are illustrative. Tap Yes, No or Undo to answer it yourself.
Works with the agents
Your manager surfaces what the specialists found.
Lumnya reads the agents’ verified facts and open recommendations and brings them into the conversation, with deep links to each agent’s page.
Wherever you are
Inside Shopify, full-screen, or on the page you’re already on.
Embedded in your Shopify admin or as a standalone dashboard — same data, same controls, same journal. And Lumnya opens as an overlay on any page, already knowing which one you’re looking at, so you never have to explain the context you’re staring at.
How you hear about it
Told once, at a sensible hour.
Urgent things email you right away. Everything else waits and arrives as one daily digest instead of a stream of pings. Web push is opt-in per device, and both lanes hold through quiet hours in your store's timezone.
Notifications
your channelsstore time
21:40
Lumnya AI · A shopper asked for a human
emailed now · waiting in Live Chats
Lumnya AI · 6 things from today
one email, 08:00 your time
Lumnya AI · Reorder suggestion ready
quiet hours — delivers at 08:00
AI-labeled · quiet-hours aware · capped per day
Your guardrails
Discreet and honest by default.
Lumnya won’t expose internal pricing rules, won’t claim it did something it didn’t, and reports errors plainly instead of pretending.
Private beta
Brief your AI manager today.
Join the private beta and be first in line when Lumnya opens to your store.