Service · Integrations

Smart API integrations between your systems.

Your CRM needs to talk to your accounting system, your ATS to payroll, your webshop to your ERP. We integrate your stack, via API, iPaaS or custom middleware, and make sure data flows reliably, even when one system is briefly down.

REST & GraphQLiPaaSCustom middlewareEvent streamingGDPR-proof
Schedule an integration call View all integrations

An integration is not an Excel export.

Most organisations that come to us share the same frustration: five or ten SaaS tools, each good at its own job, and none of them talking to each other. Someone in finance copies orders from Shopify into Exact. Recruitment manually exports candidates from Bullhorn to payroll. Customer data lives in three places at once and nobody knows which version is correct. The result: lost time, errors in the figures, and unreliable reports for management.

A good integration doesn't solve this with a nightly CSV export or a rejected Power Automate flow that nobody dares touch. It makes data flow in real time, surfaces errors before they cause damage, and stops your people wasting time re-typing data between systems. We build that layer — pragmatically, without throwing out your existing stack. See also our broader integrations overview page for the complete vendor catalogue.

What sets us apart is that we are neither blind iPaaS evangelists nor custom-build evangelists. We take an honest look at what fits: sometimes a fifteen-step Make flow is perfectly fine, sometimes you need a Kafka bus in between because the volumes simply call for event streaming. The choice depends on your context, not on what we happen to enjoy building.

Who we build this for.

Our integration projects suit a wide range of organisations, but we see a few recurring profiles. SMEs with three or more separate SaaS tools that suffer from data silos and manual handovers. Mid-market scale-ups where the original stack is starting to buckle under growth: recruitment grows by 40%, finance can no longer keep up, and the first errors on payslips start to appear.

Corporate groups with multiple business units that each run their own ERP or CRM but still want consolidated reporting. Software vendors that connect their product to their customers' stack and need a scalable integration layer rather than a custom script per customer. Agencies and consultants who want to deliver white-label integrations for their own clients without hiring an in-house development team. And finally trade associations and cooperatives that want to give their members shared data or a central portal.

Three ways to integrate systems.

The right solution depends on your volume, the complexity of the transformations, and how much control you want to keep over the logic yourself. We'll advise which one fits in the first conversation — as a systems integration specialist, we don't automatically default to one flavour.

Vendor API · direct integration

Point-to-point API integration

Two systems that talk directly to each other via their official REST or GraphQL APIs. Quick to set up, little overhead, and well suited to stable integrations with clear domain boundaries. Think Shopify → Exact, HubSpot → Mailchimp, or a payment provider to your accounting system.

RESTGraphQLWebhooksOAuth2
iPaaS · platform-driven

Integration via an iPaaS platform

For organisations with ten or more integrations, an integration platform often works better than standalone scripts. We implement Workato, Boomi, MuleSoft or Make where appropriate, with monitoring, error recovery and version control built in. Read the API vs iPaaS comparison if you're still unsure which approach fits.

WorkatoBoomiMuleSoftMakeFrends
Custom middleware · full control

Your own integration layer or API gateway

When iPaaS is too limited — complex transformations, high throughput or strict compliance — we build a custom middleware layer. With an API gateway (Kong, Tyk, Apigee), event streaming via Kafka or NATS, and a dead-letter queue for failing messages. Code in your own repository, hosted in your own cloud, with clear observability via Grafana or Datadog. This is the approach mid-market and enterprise companies grow towards as their integration landscape matures.

KongKafkaConfluentNATSDLQ

Why organisations work with us.

We think vendor-independently. We earn no partner commission, so we advise on what fits your landscape, not on what happens to earn us the most. Sometimes that's a Workato subscription we make nothing from, sometimes a custom project. The advice stays the same.

The code is yours. All custom code is committed to your repository from day one. No lock-in, no supplier contract stopping you from switching. If you're no longer happy with us in a year's time, another agency can pick up where we left off, with no reverse-engineering.

We work pragmatically. We build only what you need. No over-engineering, no ceremonial architecture diagram that nobody reads. A simple integration between two systems doesn't get a Kubernetes cluster, and a complex multi-tenant landscape doesn't get an Excel macro. The solution fits the scale of the problem.

If your webshop runs on CloudSuite, the work is mostly about prices, stock and orders between that platform and your ERP. What that involves is covered on Building a CloudSuite integration.

What you get at the end.

Not just a working integration, but everything around it so you stay in control. Our integrations are owned by the client: the code is in your repository, the credentials in your vault, and you can switch supplier at any time.

  • Working integration in productionIncluding a staging environment for testing. Runs on your own cloud (GCP, AWS, Azure) or in your existing iPaaS tenant.
  • Codebase + architecture documentationFull source code in your repository, mapping documentation, and a runbook for incidents.
  • Monitoring + alertingDashboards for throughput, latency and failure rate. Alerts to the channel of your choice when an endpoint falters.
  • Error recovery and DLQMessages that fail don't vanish into thin air. A dead-letter queue with a retry strategy and a UI for manual intervention.
  • GDPR flow documentationWhich data flows between which systems, classification per field, and where needed a DPIA contribution for your privacy officer.
  • Optional support contractMonitoring, security patches, absorbing vendor API changes, and ongoing development. Fixed monthly price with four response-time tiers.

Integrations we often build.

A cross-section from recent years. Recognise your own integration challenge? We'd be happy to talk it through.

Finance & sales

CRM to accounting

Synchronise Salesforce or HubSpot with Exact Online, AFAS or Microsoft Dynamics. Deals become invoices, customer data stays in sync, and the finance department stops double entry.

E-commerce

Webshop to ERP

Shopify, Magento or WooCommerce exchanging orders, stock and returns with your back end. Including multi-warehouse, multi-currency and VAT-correct invoicing for customers abroad.

Recruitment & HR

ATS to payroll

Bullhorn, Connexys, Homerun or Recruitee passing placed candidates on to NMBRS, Visma or Loket. Including contracts, changes and terminations. See also our page on Bullhorn integrations.

Customer communication

WhatsApp, Twilio & Slack

Automated messages to customers via WhatsApp Business, SMS via Twilio, or internal alerts in Slack and Teams. With an audit log and opt-out in line with the GDPR.

Payments

Multi-PSP payments

Mollie, Adyen, Stripe or Buckaroo connected to your accounting. At higher volumes, multiple PSPs in parallel with smart routing based on cost and success rate.

Government & data

Dutch-specific integrations

KvK Business Register, BAG, Kadaster, DigiD, eHerkenning, and Peppol for e-invoicing. Plus data warehouse integrations with Snowflake, BigQuery or Power BI.

Documents & signing

E-signing and document generation

DocuSign, Signhost, Itsme or PandaDoc for digital signing and contract handling. Including generating documents from templates using data from your CRM, ERP or HR system — no more assembling Word files by hand.

PoS & retail

Till to back office

Lightspeed, Square or MplusKASSA connected to stock, accounting and marketing tools. Real-time revenue feeds to dashboards, automatic reordering, and customer data from your till systems to your CRM for follow-up.

SSO & identity

Single sign-on and provisioning

Microsoft Entra, Okta, Auth0 or Google Workspace as the central identity provider for your entire application landscape. Including just-in-time provisioning, group mapping and automatic deprovisioning when someone leaves.

Tools and technologies we work with.

The choice of tooling depends on the project. A selection of what we use often — no dogma, just pragmatism.

iPaaS platforms

For broad many-to-many integrations

Workato for mid-market and enterprise, Boomi for traditional EAI projects, MuleSoft Anypoint for heavier implementations, Make and Zapier for light automation, and Frends for organisations that prefer a European provider. We advise neutrally based on your landscape, volume and budget.

API management

Gateway, throttling, versioning

Kong, Tyk and Apigee for exposing internal APIs to partners or a mobile app. Including rate limiting, OAuth2, key rotation and a developer portal. For simpler projects, AWS API Gateway or Cloudflare API Shield.

Event streaming

For high volume and asynchronous flows

Apache Kafka or Confluent Cloud for ordered event streams at scale, NATS for lighter pub/sub scenarios, RabbitMQ for classic message queues. Including schema registry and consumer monitoring so downstream systems don't quietly drift out of step.

Observability

Insight into every message flow

Grafana, Datadog or OpenTelemetry for metrics and traces across the entire integration chain. Distributed tracing so you can look back at where an order got stuck. Plus structured logging with correlation IDs across all systems.

Workflow orchestration

For long-running processes

Temporal or Camunda for processes that can take hours or days — for example, an onboarding flow that waits for signatures from several parties. With retries, compensating actions and human-in-the-loop steps.

AI & LLM integrations

For enrichment and classification

Integrations with OpenAI, Anthropic, Azure OpenAI or self-hosted LLMs for data enrichment, classification of incoming messages, or automatic summaries. Always with a fallback to deterministic logic, so a hallucination never breaks an integration.

How an integration project works.

No surprises afterwards. We work in fixed steps, with a working build at the end of every sprint.

1

Audit & roadmap

We map your current integration landscape — which systems, which data flows, which pain points. From that comes a prioritisation: which integration delivers the most value now, which can wait, and which first requires a data clean-up. The result is a concrete roadmap, not a PowerPoint.

2

Choosing an approach

For each integration, we advise what fits: direct API, iPaaS or custom. We look at volume, transformation complexity, vendor lock-in and maintenance burden. Honest advice — sometimes a no-code Zapier flow is enough, sometimes you need a Kafka bus in between.

3

Building in sprints

Every two weeks, a working integration or incremental extension. You test along in staging, see what is flowing, and can adjust mappings before we go to production. The code goes straight into your repository — no black box.

4

Go-live & monitoring

A phased rollout, often starting with a subset of data or a pilot department. Monitoring and alerts are active from day one. We stand by during the first sprint after go-live to respond quickly to edge cases that only surface in production.

5

Further development

Vendor APIs change, your stack grows, new systems are added. Optionally, we provide ongoing maintenance with security patches, vendor update checks and extensions on a fixed monthly basis.

Who you'll speak to

Mies Brons

You'll have the first conversation with Mies Brons, business developer. You tell us which systems aren't talking to each other and what that costs; Mies asks the questions needed to give an initial direction.

In the audit, you work with the developer who builds the integration. They map out your systems and data flows and advise which route fits.

Book an introductory call with Mies →

Compliance and data governance.

Integrations almost always touch personal data or business-critical data. That calls for more than just TLS encryption — it calls for a well-thought-out data architecture, an audit trail and clear agreements on responsibilities between supplier and client.

We design integrations with compliance built in, not as an afterthought. For regulated sectors, we provide the evidence your auditor and privacy officer need.

  • GDPRData flow diagram per integration, classification per field, data minimisation as a design principle. Contribution to your DPIA where needed.
  • DORA (financial sector)ICT risk management, incident reporting and third-party monitoring in line with the Digital Operational Resilience Act. Relevant for banks, insurers and fintech vendors.
  • NIS2 (critical infrastructure)For essential and important entities: supply chain security, log retention and mandatory incident reporting built into the integration layer.
  • NEN 7510 (healthcare)For healthcare integrations, extra controls on access, logging and data residency. Data processing agreements in order.
  • Secrets managementAPI keys and credentials in HashiCorp Vault, AWS Secrets Manager or GCP Secret Manager — never in code, never in a config file in a repository.
  • Audit logWho changed which transformation, which message took which route, which retry succeeded. Searchable and exportable for compliance checks.

Frequently asked questions.

What clients usually want to know before we start.

What does API integration cover?
All integrations between software systems where data is exchanged automatically via a standardised interface — usually a REST or GraphQL API, sometimes webhooks or event streams. In practice: your CRM pushing deals to your accounting software, your webshop syncing orders with your ERP, or your ATS passing candidates on to payroll. Government integrations such as KvK, BAG, Peppol and DigiD also fall under this.
iPaaS or custom — which should I choose?
It depends on volume, complexity and maintenance burden. iPaaS (Workato, Boomi, MuleSoft, Make) is quicker to set up and works well for relatively simple many-to-many integrations. Custom middleware pays off for complex transformations, high volumes, strict compliance, or when owning the code matters to you. A mixed landscape is also an option — iPaaS for the simple flows, custom where it counts. We wrote a detailed comparison of APIs versus integration platforms on this.
Which vendors do you integrate with?
Too many to list, but the most frequently requested are: Salesforce, HubSpot, Pipedrive (CRM), Exact Online, AFAS, Twinfield, Microsoft Dynamics (accounting/ERP), Bullhorn, Connexys, Homerun, Recruitee, HR-Office (recruitment), NMBRS, Visma, Loket (payroll), Shopify, Magento, WooCommerce (e-commerce), Mollie, Adyen, Stripe, Buckaroo, MultiSafepay (payments), WhatsApp Business, Twilio, Slack, Teams (communication), Microsoft Entra, Okta (SSO), KvK, BAG, Kadaster, Peppol, DigiD (government), Snowflake, BigQuery, Power BI (data). We can integrate virtually any SaaS system that has a documented API.
How secure is this — GDPR, encryption, auditing?
As standard: encryption in transit (TLS 1.2+) and at rest, OAuth2 or OIDC for authentication, secrets kept in a vault (never in plain text in code), and an audit log of who changed what. For every integration we produce a data flow diagram with a classification for each field (PII, financial, sensitive) — useful as input for your DPIA. Where relevant, we take DORA (financial), NIS2 (critical infrastructure) or NEN 7510 (healthcare) into account.
What determines the cost of an integration?
Three main factors: the number of systems and data flows, the complexity of the transformations (1-to-1 mapping is cheap; business logic with validation and enrichment costs more), and the level of error handling you need. A simple Shopify–Exact integration is a very different project from multi-PSP routing with fraud detection. After the audit we give you a concrete scope with a sprint budget — no open-ended hours.
How long does a project like this take?
For a simple point-to-point integration, a first working integration can be in place within a few sprints. A broader integration landscape with multiple systems, transformations and a compliance layer takes several sprints. We work in fixed sprints so you can adjust course after each one and recalibrate the scope based on what we learn.
Who owns the code?
You do. As standard, all custom code lives in your repository under your account. Credentials and API keys are stored in your own vault. For iPaaS implementations, you own the tenant and the flows — we only have access for as long as it's needed. No lock-in: you can switch implementation partner at any time without losing the integration itself.
Can you migrate from iPaaS to custom (or the other way round)?
Yes. Many clients start with a no-code tool such as Zapier or Make and, over time, run into its limits — performance, error handling, or a monthly fee that gets out of hand. We migrate step by step: first the pain points move to custom middleware or a more robust iPaaS, then the rest. The reverse works too — moving parts of a hardened custom system back to iPaaS where they're simple enough.
How do you handle multi-tenant or multi-business-unit setups?
For groups with multiple business units, or software vendors who deliver client-specific integrations, we build a layer in which tenants are logically separated. Each tenant gets its own credentials, its own mapping rules and its own logs. Configuration is done via an admin UI or declaratively in code, depending on what suits your team.
What do you do about monitoring and error recovery?
Every integration is monitored for throughput, latency and failure rate. Messages that fail to process successfully go to a dead-letter queue with a retry strategy. Critical failures trigger alerts to the channel of your choice (Slack, Teams, PagerDuty, email). For custom integrations, we also build a replay UI so your own team can step in manually without developer help.
Do you work with our own IT department?
Almost always. We hand over knowledge in the final sprints, deliver runbooks for incident response, and agree clear responsibilities. For some clients we are the main contractor, for others an extension of an internal platform team that keeps control itself. We also work with other integrations or agencies your organisation already has: no egos, just clear agreements between us.
What if the vendor API changes unexpectedly?
It's inevitable: Salesforce, Shopify and Microsoft regularly push breaking changes. For our active clients, we follow the release notes of the connected systems, test against the upcoming version in a sandbox, and patch before the deadline. For clients with a maintenance contract, this is part of the service; otherwise we warn you in good time so you can decide who fixes it.
Do you also do AI or LLM integrations?
Yes. We connect LLMs (OpenAI, Anthropic, Azure OpenAI, or self-hosted models) to business processes for enrichment, classification or automated processing. Think of categorising incoming emails, summarising contracts, or routing customer enquiries. Always with deterministic fallbacks and a human in the loop for critical decisions: an LLM should never approve invoices on its own.

Talk to us about your integration challenge.

A half-hour introduction, no obligation. Tell us which systems aren't talking, what pain that causes, and what you've tried so far. We'll give you direction — sometimes the solution is simpler than you think, sometimes it calls for a bigger step such as a data engineering platform.

Edit content