E-commerce & distributionexample use case
System integration: enter data once, have it everywhere
An order comes in from the online shop, someone retypes it into the ERP, the customer into the CRM, the invoice into accounting and a summary into a spreadsheet. The same data five times, five chances for a typo. We connect your systems so each piece of data is entered once and reaches the others on its own — mostly without AI, with AI only where rules are not enough.
Sound familiar?
- The same data is typed by hand into the shop, the ERP, the CRM, accounting and a spreadsheet — every day, by several people.
- Stock in the shop does not match stock in the ERP, and customers order items you do not have.
- Meetings are spent not on the numbers but on which spreadsheet is current and who edited it last.
- One customer exists in three systems under three different names, and one product under three codes.
How it works
1. We map where each piece of data originates
For customers, products, prices, orders and invoices we define one source of truth — the system where the record is created and changed. The others only receive it.
2. We connect systems the simplest way
Where there is an API or a ready-made connector, we use it. Where there is not, we use XML import, file exports or read the database directly. Make or n8n is often enough.
3. AI only where rules fall short
Data from emails, PDFs and free text, matching duplicate customers or products with different names. That is where a language model (LLM) helps; elsewhere it would be wasted cost.
4. Monitoring and alerts
Every transfer is logged. When something fails or a record does not match, you find out straight away — not a month later at closing.
What the system handles
- ERP and accounting: Pohoda, Money S3/S4/S5, Helios, SAP Business One, Odoo or a custom system — via API, XML import/export or the database.
- Online shops and CRMs: Shoptet, WooCommerce, Shopify, PrestaShop, HubSpot, Pipedrive, Raynet and others with an API.
- Excel and Google Sheets as an output for reports — rather than a place where data is retyped by hand.
- Syncing stock, prices, orders, customers and invoice status — in real time or in batches, as needed.
- Legacy systems without an API: scheduled file exports, direct database reads, or a custom MCP server if AI also needs to work with the data.
Human oversight and safety
- Data flows one way from a clearly defined source of truth — no two-way overwriting where nobody knows which side won.
- AI suggestions (such as merging two customers) are approved by a person first. They apply automatically only where accuracy has been proven.
- The integration has only the permissions it needs: what it may read, what it may write and what it never touches.
- Every transferred record is traceable: where it came from, when, and what happened to it.
What result is realistic
~30 h
a month less retyping between systems
For a typical company where two or three people spend 30–60 minutes a day moving data between systems, we plan for savings of around 30 hours a month. The bigger benefit is often less visible: shop stock is accurate, invoices do not go out with the wrong address, and meetings are not spent on spreadsheet versions. If classic integration without AI does the job, that is what we will propose — it is cheaper and more reliable.
The figures are indicative for typical volumes. We give a precise estimate for your company after analysing the process.
Indicative scope
- Map of systems and data flows
- €900 – €1,800
- First connection, e.g. shop ↔ ERP (2–4 weeks)
- €2,500 – €6,000
- Further systems plus AI for emails, PDFs or matching
- €5,000 – €12,000
- Operations, monitoring and connection maintenance
- from €290 / month
Prices are indicative and exclude VAT. We give an exact figure only after analysing your process — not off the cuff on the first call.
Frequently asked questions
Do we even need AI to connect our systems?
Usually not. Moving orders from the shop into the ERP or syncing stock is classic integration via an API or a connector. We add AI only where the data is unstructured (emails, PDFs) or where records need matching that do not line up exactly. If you do not need it, we will tell you.
Our ERP has no API. Can it still be done?
Usually, yes. Pohoda and Money S3 both support XML import and export, many systems allow reading straight from the database, and as a last resort scheduled file exports work. It is slower and more fragile than an API, so we design it with checks and alerts.
Why not just Make or Zapier?
Often we do use Make or n8n — for simple flows they are the right tools. We write our own code when the number of scenarios grows, when errors, retries and state need handling, or when data must not pass through yet another third party.
What about the duplicate customers and products we already have?
They need cleaning up before the systems are connected, otherwise the mess just spreads. Rules handle exact matches; for ambiguous ones (a name spelled differently, an old address, a different code) AI proposes a match and a person confirms it. With thousands of records that saves days of manual work.
How long does it take?
Connecting two systems with APIs usually takes 2–4 weeks. More systems, a system without an API or cleaning up old data stretches that to 6–10 weeks. We give a firmer estimate after the systems map.
Related service
Connecting AI to your systems
We build this solution as part of our service „Connecting AI to your systems“.
We connect AI to the systems you already useRead more
- How to connect your ERP, CRM and Excel so nothing is entered twiceSystem integration without double data entry: setting a source of truth, choosing a route (API, Make, file import), coping without an API, and where AI helps.
- MCP: the protocol that finally made AI a usable toolThe Model Context Protocol is roughly to AI what USB was to peripherals. An explanation without the marketing — and what follows from it when you are considering an integration.
- How to automate order processing from email: rules, Make or an AI agent?Orders arrive by email as text, PDFs and spreadsheets, and someone retypes them into the ERP every day. Three ways to automate it, where each one breaks — and what you need ready before you start.
More solutions
DistributionReporting automation: a management report that builds itself
~10 h · saved on reporting every week
WholesaleAutomating email orders into your ERP with AI
~70% · of orders handled without a human
ManufacturingCompany document search — an answer with its source in seconds
2–4 h · saved per person per weekIf you can say how many systems someone at your company types the same order into, we have a number to start from.
You do not need to know whether you need AI, automation or a new system. Show us the process that slows you down — we will tell you what can be automated and whether it pays off.