Skip to content
profflowLog inINSTALL FREE

AI operator

Your numbers.
Now, explained.

Ask about profit, costs or what changed. Get a clear answer grounded in the figures already inside your dashboard.

Built for Shopify · Your figures. Your control.

Profflow AISample store · try it

You askedWhy did profit fall?

ProfflowAnswered by rules from the dashboard figures · profit

Net profit this month is €2,521 on €7,940 of revenue — a 31.8% margin. You made €2,521 after every cost.

Last 7 days against the 7 before:

  • Profit €435 vs €653 (−33.4%)
  • Revenue €1,609 vs €2,046 (−21.4%)
  • Orders 14 vs 18 (−22.2%)

What moved it, largest first:

  • Revenue fell by €437 (−€437 on profit)
  • Cost of goods fell by €116 (+€116 on profit)
  • VAT fell by €79 (+€79 on profit)

So the short answer: revenue fell by €437. That is the line to look at first.

These are demo figures — connect your Shopify store to see your own.

The sample store, one month of it. The answers are worked out by the engine the dashboard uses, not by a model reading a chat; on Pro a model handles open questions on top of the same figures.

01 / The starting point

Not an empty chat.
Your business context.

The assistant reads the context Profflow has already calculated for your current dashboard. The same figures on screen become the basis of its answer.

  • Orders & revenue

    The same connected totals displayed in the dashboard.

  • Refunds & fees

    Posted deductions, kept in the period where they happened.

  • Cost model

    Goods, shipping, packaging, ads, fixed costs and VAT.

  • Time windows

    Last 7, previous 7, last 30 and previous 30 days.

  • Products

    Tracked product revenue, units and clearly labelled estimates.

  • Destinations

    Country mix and VAT context where the data is available.

02 / Start here

Ask the questions
operators ask every day.

Four prompts are available immediately. Pro and Business can ask broader questions through the model-powered assistant.

  • “Why did profit fall?”

    Compares the last seven days with the seven before and ranks the revenue and cost movements behind the change.

  • “What are my biggest costs?”

    Orders each available cost line by size, share of revenue and recent movement.

  • “What changed this week?”

    Reads revenue, profit, orders, refunds and margin across two equal periods.

  • “Which products bring the most profit?”

    Ranks measured product revenue and labels contribution as an estimate whenever product-level cost is missing.

03 / Guardrails

Useful because
the ground is visible.

  • One dashboard context

    The assistant receives the figures already rendered for the signed-in business instead of building a second version of profit.

  • Deterministic fallback

    Supported questions have a rule-based answer underneath the model, so a model outage does not turn into a guessed financial reply.

  • Missing stays missing

    When the context does not contain the number required for an answer, Profflow says what is missing instead of filling the gap.

  • Estimates stay labelled

    Demo figures, incomplete cost setup and estimated product contribution are named inside the answer.

04 / How it answers

One context.
Two answer paths.

Basic answers are deterministic. Smart answers add a model that reads the same context, with the deterministic engine underneath it.

  • 01 · Dashboard context

    Totals, costs, periods, products and destinations are assembled from the current view.

  • 02A · Rules

    Supported financial questions are calculated directly and remain available as the fallback.

  • 02B · Model

    On Smart AI plans, the model handles broader language using the same supplied facts.

  • 03 · Answer

    A concise explanation, with estimates and missing context named inside it.

05 / Limits

What it does
not pretend to be.

  • No tax filing

    Profflow prepares VAT-aware figures; it is not an accountant and does not submit returns.

  • No invisible actions

    The AI reads business context. It does not silently change prices, orders, spend or customer records.

  • No perfect attribution claim

    Channel attribution and business contribution remain distinct signals.

  • No product-margin fiction

    If true SKU or ASIN cost is missing, contribution is presented as an estimate rather than a measured margin.

About the
AI operator.

Can’t find the answer to your question?
Contact us by clicking here.

  • For supported financial questions, the deterministic answer engine returns an answer from the same dashboard context.

  • The system is designed to answer only from supplied figures and to stop when required context is missing. As with any model, Profflow keeps evidence visible so the operator can verify the conclusion.

  • Growth includes deterministic answers. Pro and Business add the model-powered assistant on top of the same figures.

The figures are already there. Now make them answer.

Open an AI inside the dashboard and ask about profit, costs, products or what changed.