Connect your systems: a guide to integrations
When systems don't talk to each other, the same data gets keyed in several times. That's duplicated work, errors waiting to happen — and time that disappears.
Quick overview
An integration connects systems you already have so data flows automatically instead of being keyed in several times across your web shop, accounting system and CRM.
How to get started:
- Map where the same data is entered twice and causes errors and delay.
- Choose a concrete connection via API, e.g. Fortnox ↔ HubSpot or Business Central ↔ website.
- Build it with error handling, monitoring, documentation and idempotency.
Bear in mind: you don't need to switch systems — a good integration keeps what works and is only finished when it keeps working and makes itself visible when it doesn't.
Most companies already pay for good systems. The problem is rarely the tools — it's that they aren't connected. An order is typed into the web shop, then again into the accounting system, then again into the CRM.
Every such hop is a manual step that someone has to remember to do, find the time for and get right. It rarely registers as a big problem, because it's spread across the day and across several people. But it adds up to something significant: time spent moving data instead of using it, and a constant risk that the figures in two systems don't match. An integration removes the hop itself — data that already exists in one system is transferred automatically instead of being copied out by hand.
What does it really cost?
- Duplicated work — the same task is done several times.
- Errors — every manual transfer is a chance to type something wrong.
- Delay — data goes stale before it reaches the next system.
- Worse decisions — no one has the full picture in one place.
Common connections
Many of the most valuable integrations are undramatic — they simply save time every day.
The key point is that you don't have to replace anything to get there. The whole point of integration is the opposite — to keep what works and get the systems talking together. Most systems with an API can be connected, and the most valuable connections rarely involve new functionality, but rather removing a recurring manual step.
What makes a good integration?
A connection that works during the demo isn't the same as a connection that holds up in production for a year. The difference lies in what happens when reality doesn't look like the plan — when a system is down for a moment, when a field changes format, or when the same job happens to run twice. That's where a robust integration differs from a brittle script.
- Error handling — what happens when something goes wrong? Is it logged and alerted, or does the data disappear silently?
- Monitoring — do you notice if the flow stops, or is it only discovered when a customer gets in touch?
- Documentation — does anyone other than the builder know how it works and why it was built the way it was?
- Idempotency — the same data isn't sent twice on a re-run, so an order or invoice doesn't end up duplicated.
Common integrations
The most valuable connections are rarely advanced — they simply remove a recurring manual step. The table shows where the same data would otherwise be entered several times, and what you gain by connecting the systems.
| System A | System B | What you gain |
|---|---|---|
| Web form | CRM (e.g. HubSpot) | New leads land directly with the salesperson without anyone copying them out by hand. |
| Order (web shop) | Fortnox | The order is booked automatically instead of being typed in a second time. |
| CRM | Invoicing | A won deal becomes an invoice basis without duplicate registration. |
| Shopify | Business Central | Stock and order data are kept in sync between shop and business system. |
| Google Ads | CRM | You see which ads actually led to deals, not just to clicks. |
Common pitfalls
Most integrations that grate have the same causes. Watch out for these — they determine whether the connection holds up over time or becomes a silent source of errors.
- No error handling — when a call fails, the data disappears silently instead of being logged and alerted.
- No monitoring — the flow stops, but no one notices until a customer gets in touch.
- Missing idempotency — a re-run sends the same order or invoice twice.
- Undocumented — only the person who built the connection knows how it works, and can't explain it to the next person.
- Too tight a coupling — a small field change in one system breaks the whole flow, because no intermediate layer catches the difference.
Want to see what this looks like in practice? Explore our solutions, see case studies and platforms we've built, or get three free suggestions.