Solution · Reports and dashboards
Automated reporting for SMEs: code calculates, AI explains, you approve
Every week or month, the agent reads your exports (Excel, CSV) or your software’s API, calculates your KPIs, flags what’s drifting and writes a clear report for management, your clients or your board. You review it, it goes out.
- Figures calculated by code, never guessed
- Reports in French, German or English
- Human approval before anything is sent
What an automated reporting agent does
In many small businesses, the monthly report is still built by hand: export the accounts, paste into Excel, redo the charts, write three sentences for management. By the time it’s ready, the problems it shows are often weeks old.
Automated reporting always follows the same pipeline. The agent collects your data: we read your exports (Excel, CSV) or your software’s API (accounting, CRM, online shop, campaigns). Then code calculates every KPI: revenue, margin, receivables, billable hours. Thresholds agreed with you flag anything unusual. Only then does AI explain: it writes a plain-language summary from the results, without ever doing the maths itself.
The result: an up-to-date SME dashboard and AI business reports that take two minutes to read, in French, German or English. A person approves each one before it goes to management, a client or the board. And if a KPI crosses a threshold mid-month, an alert goes out straight away, without waiting for the month-end close.
The problem
Where the time goes today
The monthly report eats hours
Exports, copy-paste, formulas, charts to redo: every month, someone rebuilds the same document by hand, with the risk of a copying error.
Problems surface too late
An unpaid invoice getting older, a margin slipping, a campaign running out of steam: it often shows up at month-end, when it’s late to react.
Figures without explanations
A table of 40 KPIs doesn’t say what changed or why. Management wants three clear sentences and the decisions to make.
Interactive demo
Try the reporting agent
Pick a fictional company and the report language, then run the agent: it collects the data, calculates the KPIs in front of you, spots anomalies, writes the summary and draws the charts. Change a figure in the text and the check flags it.
Reporting agent
Gathers your figures, calculates, spots the gaps and writes the report. You approve before it goes out.
Sources Input
- CSV Accounting · ledger —
- CSV Open receivables —
- Excel Timesheets —
- API Tax files —
- CSV September report —
Calculations Code
- Fees
- Costs by line
- Operating margin
- Cost increase
- Receivables > 30 days
- Concentration
- Billable hours
- 2025 tax returns
Anomalies Code
- Receivables > 30 days · change
- Margin · change
- Billable-hours rate
- Tax returns received
Log
Everything runs in your browser: nothing is sent or stored. Simplified demo: the figures are calculated in your browser from fictional data, and the summary is a template filled in with those figures. In production, a language model writes from the calculations, and someone on your team signs off.
I want this for my businessWhat’s happening, step by step
-
The figures come in Input
The agent reads your exports (Excel, CSV) or your software’s API: accounting, CRM, online shop, advertising.
-
Code calculates Code
Revenue, margin, change on last month: fixed formulas, like a spreadsheet. No AI here.
-
Gaps stand out Code
Thresholds agreed with you flag anything unusual: receivables up, margin down, stock running low.
-
The AI writes the summary AI
A language model receives the calculated results and explains them in plain words, in the reader’s language. It explains; it doesn’t calculate.
-
The report takes shape Output
Key figures, charts and points to watch: a document you can read in two minutes, online or as a PDF.
-
You approve Human
You read it, correct it if needed, and the report goes to management, the client or the board.
Run the demo: each step lights up here.
ModelApertus 70B · Swiss open source
Data processedA server in Switzerland or yours
Picture a financial controller who, on the first morning of the month, has already gathered all your figures, calculated every KPI, spotted what’s off and written a clear summary. Code calculates, AI explains, you decide.
Words to know
- KPI
- Key performance indicator: a figure you track over time, such as margin, receivables or conversion rate.
- Alert threshold
- The limit agreed with you beyond which a figure is flagged, for example “receivables up 20% in a month”.
- Dashboard
- A page that brings your KPIs together and updates itself.
Why it’s safe
- Figures are computed by code from your data, not by the AI.
- Any number in the text that does not match the calculation is flagged.
- Nothing goes to management until a person has reviewed and approved it.
- You choose where the AI and your data run.
The agent’s pipeline
- Input Sources Excel or CSV exports, your software’s API (accounting, CRM, shop, advertising). Read-only, on a schedule.
- Code Data checks Complete periods, duplicates, totals reconciled with the accounts. A missing source stops the report.
- Code KPI calculation Python and SQL: sums, ratios, change on last month and on budget. Versioned, tested formulas.
- Code Anomaly detection Two-level thresholds agreed with you, plus unusual moves compared with history.
- AI Writing The model only receives the computed facts (JSON) and writes the summary in the chosen language.
- Code Text check Every number quoted in the text is compared with the calculations; a mismatch is flagged before sending.
- Human Approval and delivery Review, then delivery by email, as a PDF or on a web dashboard.
AI reads and writes. Code checks. Humans decide.
Stack
Guardrails
- Code calculates, AI explains: the model only receives results that are already calculated, never columns to add up.
- Number check: every figure quoted in the text must match the calculations, or the report is flagged before sending. Try it: change a figure in the demo.
- Checked sources: an incomplete period or totals that don’t match the accounts, and the report stops with an alert.
- Read-only access to your software, and a log of every run.
- Human approval by default: automatic sending is decided report by report.
The facts passed to the AI (JSON)
{
"report": "monthly_management",
"company": "Alvélia Fiduciaire Sàrl (fictional)",
"period": "2026-10",
"language": "en",
"facts": {
"fees_chf": 168300,
"fees_change_pct": 6.1,
"operating_margin_pct": 15.3,
"margin_change_pts": -4.7,
"cost_drivers": [
{
"line": "temporary staff",
"change_chf": 9800
},
{
"line": "licences",
"change_chf": 4200
}
],
"receivables_over_30d_chf": 37750,
"receivables_change_pct": 38,
"billable_rate_pct": 64.9
},
"alerts": [
{
"rule": "receivables_over_30d_change > 20%",
"value": 38,
"level": "alert"
},
{
"rule": "margin_change < -2 pts",
"value": -4.7,
"level": "alert"
},
{
"rule": "billable_rate < 65%",
"value": 64.9,
"level": "watch"
}
],
"summary": {
"status": "awaiting_approval",
"numbers_checked": true
}
}Choose the engine
How we set it up
- Day 1 Free audit 30 minutes to list your sources, your KPIs and who gets what.
- Weeks 1–2 Prototype on your figures A first report built from your exports of the last few months; every formula is checked with you.
- Weeks 2–4 Go-live Schedule, thresholds, recipients and approval rules.
- Every month Care Source monitoring, threshold tuning and new KPIs when you need them.
What we need from you
- An export of your data, or read access to your software
- Your current report, even a hand-made one, or your list of KPIs
- Who approves the report and who receives it
What we set up
Reports that prepare themselves, and that you review
Your KPIs, your thresholds, your recipients. Not a generic template.
Monthly management report
Revenue, margin, costs, receivables, billable hours: a report ready on the first working day of the month, changes explained. A natural fit for accounting firms.
Client reports for agencies
Campaign results, cost per order, newsletter: one report per client, in their language, reviewed before it goes out. See marketing agencies.
SME dashboard
A web page that is always up to date, with the 5 to 10 KPIs that really matter to you. Nothing more.
Anomaly alerts
Receivables up, margin down, stock running low, conversion dropping: an alert goes out as soon as a threshold is crossed, with the figure and the rule.
Board summary
One page in plain language for a board or partners: what changed, why, and what needs deciding.
Clean, reusable data
The consolidated figures stay available in Excel or CSV: clean data, ready to import into your accounting software.
How we work
How it goes
-
Free audit
In 30 minutes we look at your data sources, your KPIs and who needs to receive what.
-
Prototype on your figures
A first report built from your exports of the last few months, every formula checked with you.
-
Go-live
The report runs on your schedule, with your approval rules. We aim for a first automation in service 2 to 4 weeks after the audit.
-
Monthly care
We monitor the sources, fine-tune the thresholds and add KPIs when needed.
Our commitments
Correct figures, under your control
Code calculates, AI explains
Every figure comes from a calculation you can check. The AI writes the commentary; an automatic check makes sure every number it quotes matches the calculation.
You choose where your data is processed
Leading models (GPT, Claude, Gemini, Mistral…) or, for sensitive data, open-source models such as Apertus, the open Swiss language model built by EPFL, ETH Zurich and CSCS, hosted in Switzerland or on your own infrastructure.
Built with Swiss data protection in mind
A data processing agreement (FADP art. 9), only the data the task needs, Swiss hosting available, read-only access to your software.
Frequently asked questions
What is automated reporting?
It’s a report that prepares itself, on a fixed date, from your own data: an agent collects the figures, code calculates the KPIs, AI writes the summary and a person approves it before it’s sent.
Can the AI get the figures wrong?
The figures don’t come from the AI: code calculates them, as a spreadsheet would. The AI writes the commentary from the results, and an automatic check makes sure every number it quotes matches the calculation. A person reviews it before it goes out.
Which KPIs should an SME track?
The ones that trigger a decision, usually 5 to 10. Accounting firm: fees, margin, receivables, billable hours. Agency: results and cost per client. Online shop: sales, margin, conversion, stock. We choose them together during the audit.
Which software does it work with?
With most software that offers an export or an API. We read your exports (Excel, CSV) or your software’s API: accounting, CRM, online shop, ad platforms, website analytics. If a source has neither, we find a way during the audit.
Our figures are confidential. Where are they processed?
Wherever you decide: a leading model under a data processing agreement, or an open-source model such as Apertus, hosted in Switzerland or on your own infrastructure, depending on how sensitive the data is. It’s all built with Swiss data protection (FADP) in mind. See our guide ChatGPT at work and data protection.
How long does it take to set up?
We aim for a first automation in service 2 to 4 weeks after the audit, depending on how many sources we connect and how clean the exports are.
How much does it cost?
On quote: the price depends on the number of sources, KPIs and reports. The audit is free; you then get a clear quote with a set-up cost and a monthly fee. Request a quote.
NotHuman
How many hours does your monthly report cost?
In a free 30-minute audit, we look at your sources and your KPIs, and propose a concrete plan.