Job 05 · Get answers without an analyst

Ask your store a question.

“What sold best last week? Why did revenue dip on Tuesday?” Plain language in, plain language out — with the working shown and the action one approval away.

Today, without it

Ten dashboards, no answers.

Analytics tells you what happened. The why — and the what-now — is still homework.

A number moved and nobody can say which driver moved it
The report you need is three exports and a pivot table away
Insight on Monday, action on Thursday, relevance gone by then
Charts that describe the problem instead of proposing the fix
Questions you don't ask because asking them costs an hour

Ask in plain language

The dashboard answers back.

Lumnya reads your live store data and answers the question you actually asked — no query language, no chart-building, no waiting for someone to run it.

lumnya console confirm-gated

>

… reading order #1043 · drafting refund + apology email

preview

Refund €42.00 (full) — order #1043

Email: “We’re sorry about the delay — your refund is on its way.”

Cancel Approve refund

Refunded & email sent · logged 14:32 · approved by you

It shows its work

An answer you can check.

Revenue Detective decomposes a move into quantified drivers — volume, basket size, refunds, mix — so “sales dropped” becomes a list of reasons with numbers next to them. Every specialist attaches its reasoning to what it reports.

>how are sales this week?

Revenue

€12.4k

+8%

Orders

214

+5%

AOV

€58

+2%

Up 8% on last week, driven by the hoodie restock. One watch-out: Oxygen Pro margin dipped after Tuesday’s discount code.

Show low-stock productsReview that discount

From answer to action

Then tell it to fix the thing.

Prices down to the right variant, refunds, discount codes, customer emails, settings, triggers and FAQ — you ask, Lumnya shows you the exact change first, and nothing happens until you approve it.

multi-model engine

routine workfast models
revenue-critical momentsstronger models
high-stakes reasoningfrontier models

How it runs

Four steps, no project plan.

01

Ask

In plain language, the way you'd ask a colleague who knows the store.

02

Read the reasoning

The answer arrives with its drivers and its working attached.

03

Preview the change

If it proposes an action, you see exactly what would change.

04

Approve — or don't

One tap. Approved actions execute and land in the audit trail.

The honest edges

What it won't do

An honest read on the analysis you get.

Change a price, issue a refund or email a customer without your approval
Answer from data it doesn't have — it tells you what's missing instead
Pretend to a certainty it can't support
Reach outside your store's own isolated data to answer you
Make silent changes — every action is logged with what changed

Why it matters

Not a chart. An answer, and then a fix.

The gap between knowing and doing is where most stores lose the week.

Straight answers

Questions merchants ask about this one.

No. It reads live store data to answer, and it can act on what it finds — price edits down to the variant, refunds, discount codes, emails, settings — with every consequential action previewed and confirm-gated.

It tells you what it doesn't have rather than filling the gap with something plausible. Low confidence surfaces as low confidence.

Your store's own isolated data. Nothing is pooled across merchants, and your data isn't used to train foundation models.

The other four

Something else eating your day?

Private beta

Hand over this job first.

Join the private beta and be first in line when Lumnya opens to your store.