Hours of re-typing
Someone copies the same order into the CRM, the operations sheet and the accounts package. That is a salary spent moving data sideways.
✦ Integration
Software integration services that stop the double entry
The same order typed into three systems is not a data problem, it is a margin problem. We connect the tools you are keeping so information moves once, automatically, with an audit trail.
✦ The cost
What disconnected systems are really costing you
Errors nobody notices for a month
A quantity mistyped at step two surfaces as an invoice query at step nine, by which time three people have to unpick it.
Nobody trusts the numbers
Two systems disagree, so every meeting starts by arguing about whose figure is right instead of what to do about it.
Customers wait for answers
'Let me check with the warehouse and call you back' is a phrase that only exists because two systems cannot see each other.
✦ Approaches
Four ways to connect systems
Which one applies depends on what your systems expose, not on what we prefer to build.
Direct API integration
Where both systems have a decent API. Cleanest, fastest and cheapest to maintain - always the first thing we check for.
Middleware layer
A small service between several systems that handles mapping, retries and failures. Sensible once three or more tools are involved.
Scheduled data sync
Batch transfers on a schedule where real time is not required. Simple, reliable and often perfectly adequate for finance data.
Legacy connectors
For systems with no API: database-level integration, file drops, or a careful automation layer. Not elegant, but it removes the typing.
✦ Coverage
Systems we connect regularly
Accounting
Xero, Sage, QuickBooks, Microsoft Dynamics - invoices, credit notes, payment status and nominal coding.
CRM
Salesforce, HubSpot, Dynamics and Pipedrive, so a won deal creates the job rather than an email asking someone to.
ERP and MRP
Stock, works orders, bills of materials and purchasing, including older on-premise systems.
Warehouse and logistics
WMS platforms, courier APIs, tracking updates and proof of delivery flowing back automatically.
Payments
Stripe, GoCardless and card gateways, reconciled against invoices rather than reconciled by hand.
Microsoft 365 and Google
Single sign-on, calendars, SharePoint documents and automated email or Teams notifications.
✦ Reliability
What we build in so integrations don't fail silently
Most integration horror stories are not about the happy path. They are about what happens when something breaks at 2am.
- Automatic retries with sensible backoff when the far end is down
- A dead-letter queue so nothing is lost when a record cannot be processed
- Alerting to a human - email or Teams - when something needs attention
- Idempotent handling, so a repeated message never creates a duplicate order
- A log of every record moved, with the payload, for when finance asks what happened
- A dashboard showing sync status so you can see it is working without asking us
✦ Scoping
We start by mapping the data, not the tools
Before writing anything we map the fields: what a 'customer' means in each system, which one is the source of truth, what happens when they disagree, and which records are already duplicated.
That mapping is the deliverable that saves the money. Most failed integrations fail because two systems held different definitions of the same word and nobody wrote it down.
✦ Questions we get asked
Integration questions
Our old system has no API. Can it still be integrated?
Usually yes. Depending on the system we can read from its database, exchange files on a schedule, or build a careful automation layer. It is less elegant than an API, but it still removes the re-typing.
Is real-time sync always better?
No. Real time costs more to build and monitor. Finance data overnight is often fine; stock levels usually are not. We decide field by field rather than applying one rule to everything.
What happens when a sync fails?
It retries, and if it still cannot succeed the record goes to a dead-letter queue and a human is alerted. Nothing is silently dropped, and the log shows exactly what was attempted.
Can you use Zapier or Make instead of building something?
Sometimes, and we will say so. For low volume with simple mapping they are cheaper. Once volume rises, mapping gets complex or failures matter financially, a purpose-built integration is more reliable and often cheaper to run.
Who maintains it when a vendor changes their API?
We do, under the monthly retainer. Vendors deprecate endpoints with little notice, and keeping ahead of that is part of the service rather than a surprise invoice.
Tell us what isn't working. We'll tell you what we'd build.
Thirty minutes. Bring the process, the systems it touches and the deadline. You'll leave with an approach and an honest answer on whether custom software is the right call.
No pitch deck · We map your process · You own the software