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.