AI Reporting & Dashboards
Custom dashboards and plain-English answers built on your ERP, accounting, CRM and commerce data. Including the numbers your current software cannot show.
What breaks today
Your ERP knows gross margin. It does not know the freight you absorbed, the rebate you will pay in Q4, the returns that came back, or that this customer pays weeks past terms. So "which customers actually make us money?" becomes a week-long spreadsheet project — and goes stale the day it is done.
We connect the systems, define the numbers with you once, and build views that update themselves. Then an agent writes the Monday brief so you do not have to read six dashboards.
Most reporting arguments are not about the data. They are about whose definition of margin is right. We settle that first, write it down, and build everything on top of it.
Each one answers a question that crosses systems, which is why your ERP cannot produce it on its own. The sketches beside them are shapes only — layout, not data.
The problem
Your ERP shows gross margin at list. It does not subtract freight, rebates, returns, payment terms and the cost of the account.
What you see
The problem
Cash, inventory and receivables live in three systems, so nobody sees the cash cycle in one place.
What you see
The problem
Every new question becomes a spreadsheet project.
What you see
The problem
You find out an account is slipping when the order does not come.
What you see
Open-order reports are a list, not an answer.
Commission spreadsheets are not performance reports.
The numbers exist; nobody has time to read six dashboards.
No report to pick, no filters to set. You type the question. It answers from the data it can see, and shows you the records it used.
You
Why did gross margin drop in March?
Agent
Three things, in order of impact: (1) freight on two large customers was not recharged after the carrier rate change; (2) a price-list update for one product family went out late; (3) a higher share of orders came from your lowest-tier customers. Here is each one with the orders behind it.
Illustrative example. Answers always show the records they came from.
A dashboard nobody opens is worse than no dashboard — it looks like the question got answered. We track which views get used and which questions get asked, and we retire what nobody reads.
Reporting agents read and write nothing back into your systems. The only approvals that matter here are about the numbers themselves.
| Runs on its own | Asks a person first | |
|---|---|---|
| Dashboards | Refresh automatically from your systems | Changes to a metric's definition |
| Ask your data | Answers with sources shown | Anything outside the data it can see — it says so instead of guessing |
| Monday brief | Written and sent on schedule | Distribution list changes |
| Alerts | Flags against thresholds you set | New alert rules |
A build fee, then a fee tied to what the reporting actually delivers.
1. Build fee
A build fee to connect your systems, agree definitions with you and build the first set of views.
2. Per-outcome fee
After the build, you pay per unit of work the agent actually completes. The unit is agreed up front and it is something you can count.
| Agent | You pay per |
|---|---|
| Dashboards | monthly fee per live view set |
| Ask your data | per answered question above an included monthly allowance |
| Monday owner brief | per brief delivered |
Reporting agents read from your systems and never write to your ERP. Each connection is its own scoped login, so you can see exactly what it can reach and switch it off in one place.
We set up a small store in your cloud or ours and refresh it on a schedule you choose. If you already have somewhere the data lands, we use that instead of adding another one.
Margin, landed cost, cash cycle and active customer are agreed with you, recorded in plain language and version-controlled — so when a number changes, you can see what changed about the definition.
Views open in a browser from anywhere. Access follows your own roles: a rep sees their accounts, the owner sees everything.
Ask-your-data answers link back to the orders, invoices and lines they were built from, and refuse to answer when the data is not there rather than filling the gap.
There is shared scaffolding underneath — connections, refresh, logging, access — but the views, the definitions and the rules behind them are built around how you already run. We fit your process rather than asking you to fit ours.
Plus bespoke and legacy systems — if your team can log into it, we can usually work with it.
Each agent gets its own scoped login with only the permissions its job needs. No shared admin accounts.
Each action is recorded with what the agent saw and why it acted. You set the approval thresholds — dollar amounts, new customers, price changes.
We never use your data to train models. NDAs are standard on every engagement.
Most owners start with one view, then put an agent on whatever it exposes.
From the PO in your inbox to the cash in your bank — quoting, orders, suppliers, invoices.
Find the revenue already sitting in your customer list — and follow up on every lead.
Questions & Answers
Straight answers on what is different, what it needs and who sees what.
Not sure whether reporting or an agent is the first move? See how a build runs or read the engagement and pricing model.
We will show you what it takes to answer it every week, automatically.