arades service · Cluster specialization · API-first
ERP, eCommerce, logistics, banking — Microsoft Dynamics 365 and Microsoft 365 rarely stand alone. We build interfaces that survive Microsoft Release Waves: documented, tested, with built-in error handling and monitoring. Live references from production projects with Oxaion, Shopify, Amazon, and GLS.
Managing Directors · Owners
Duplicate entry, Excel bridges, and manual data transfers cost real money — usually invisibly in day-to-day operations. A clean interface typically replaces the equivalent of 1–3 FTEs in manual data maintenance and measurably speeds up order and shipping processes. We deliver the integration business case in euros: effort, follow-up costs, hours saved — decision-ready for the board.
Department Head · Business Unit
Sales, service, and logistics each work with their own data states — not out of ill will, but because the bridges aren't there. In the discovery spike (price on request) we deliver a clean data map: source systems, master-data ownership, mapping table, consolidation path. Documented outcome report for steering committee and procurement — before the first connector gets built.
IT Lead · CIO · Solution Architect
Patterns: REST with OAuth 2 and API Management facade, SOAP legacy with XML mapping, event-driven over Azure Service Bus and Event Grid, custom connectors for Power Platform, Dataverse Web API, and Microsoft Graph. Idempotency, retry with exponential backoff, dead-letter queues, monitoring with Application Insights. Direct conversation with the specialist — no sales filter.
For Managing Directors · Integration ROI in euros
We deliver the integration business case on a single page: FTEs saved, duplicate entry avoided, accelerated cycle times, risk costs from data silos. Plus an estimate of effort and follow-up costs. What the CFO sees: a quantified investment with a clear payback. Fixed-price entry via discovery spike — no open-ended hourly billing, no unpleasant surprises.
For Department Heads · data map upfront
Discovery-spike delivery: source-system inventory, mapping table (fields, types, mandatory flags), master-data ownership per domain, consolidation roadmap with quick wins. Plus a business-case skeleton for CFO and procurement. With this, you talk to other department heads not about Excel problems but about a documented architecture decision — ready for the next steering meeting.
For IT leads · 45-min architecture conversation
Topics: point-to-point vs. event-driven architecture, Azure Service Bus (queues, topics, sessions), Event Grid for cross-domain events, OAuth 2 / OIDC token flows, API Management as a facade, idempotency and retry with exponential backoff, dead-letter queues, Microsoft Graph vs. Dataverse Web API, custom connector vs. Logic Apps. Direct conversation with the integration architect.
What this service is
An interface isn't finished when it transmits data for the first time. It's finished when it still runs after three Microsoft Release Waves, five API version jumps on the third-party side, and a change of authentication model — and when you know immediately if it ever doesn't.
We've been building interfaces for over 20 years. In the early days these were XML web services and file drops over FTP; today they're REST APIs with OAuth 2 authentication, event webhooks, and Microsoft Graph. What hasn't changed: a production interface needs more than the happy path. It needs error handling for third-party API outages, idempotency so retries don't create duplicates, retry strategies with exponential backoff, monitoring with escalation capability, and documentation that someone other than the original developer can read.
That's exactly what we deliver. An arades interface ships with an architecture document, mapping table, automated tests, monitoring dashboard, and handover session. We build both classic point-to-point connections and event-driven architectures with a message bus, depending on what the requirement supports. And we're honest when an existing architecture is the actual problem — sometimes an API facade in front of a legacy system is the better investment than the twelfth point-to-point connection.
Six integration areas
SAP, Oxaion, Microsoft Business Central, other mid-market ERPs. Master-data sync for customers, items, terms. Order handoff, invoice status return, posting-record export. Exception queue with escalation logic. Oxaion has been running live for years.
Shopify (multi-shop), WooCommerce, Magento, Salesforce Commerce Cloud. Inventory push, order pull, shipping status return with tracking. Multi-shop integrations with centralized stock logic. Webhook-based with fallback polling.
Amazon Reseller API (multi-marketplace DE, EU, and worldwide), eBay Trading and Sell API, Otto Market, idealo Direktkauf. Listing sync, order pull, status push, returns handling. High share of idempotency logic due to aggressive retry policies on the marketplace side.
GLS, DHL, DPD, UPS, Hermes. Label generation, tracking numbers, status webhooks. Multi-carrier routing by weight, country, service level. GLS is the standard integration; further providers follow the same architecture — swappable as a plug-in.
DATEV (posting-record export, master-data sync), lexoffice, Stripe, PayPal, Klarna. Payment receipt reconciliation, dunning status update, posting-record generation. PSD2-compliant for banking APIs where required.
Microsoft Graph for Microsoft 365 data — Teams, SharePoint, Outlook, Entra ID. Plus any custom APIs over REST, OData, or SOAP. If a platform has a documented interface, we connect to it — in 20 years, we've encountered very few that couldn't be done.
Architecture principles
The API is the contract — not the implementation. We design the mapping and endpoint logic first, then the code structure. Direct database access or "we'll fix it with SQL later" is the exception, not the rule. A cleanly defined API stays readable after the next Release Wave.
Not every interface needs to be event-driven. If a third-party system offers webhooks and the load justifies it, we build event-driven with a message bus. If a nightly file drop is functionally enough, we build a nightly file drop. Architecture follows requirements, not fashion.
Processing a message twice must not create two records. Every arades interface uses correlation IDs, deduplicated processing queues, and upsert logic instead of blind insert. That's how we survive aggressive retry policies from marketplaces and duplicate webhook deliveries without data chaos.
Exponential backoff with configurable upper bound, dead-letter queue for non-recoverable errors, alert escalation by mail, Teams, or PagerDuty. Monitoring dashboard shows throughput, latency, error rate per endpoint. If an interface fails, you know before your end customers do.
Three delivery models
Not every interface needs the same contract model. A cleanly scoped requirement lives in fixed price. An evolving integration landscape lives in sprint logic. An existing interface lives in the Care model.
When: Cleanly scoped requirement with clear scope — e.g., "orders from Shopify into Dynamics 365 with status feedback". Duration: 4 to 8 weeks from kickoff. What you get: Fixed-price offer after requirements workshop, fixed delivery dates, acceptance tests against the spec. Predictable, plannable, billable.
When: Requirements still evolving — multi-shop build-up in growth, marketplace expansion, new third-party integration with unclear scope. Duration: Two-week sprints, cancelable anytime. What you get: A production-capable increment every two weeks, transparent hourly billing, shared backlog grooming.
When: Interfaces are live and need maintenance — Release-Wave adjustments, third-party API version jumps, smaller extensions. Duration: Monthly flat rate, minimum 12 months. What you get: Monitoring, SLA-driven bug fixes, preventive Release-Wave testing, hour quota for ongoing development.
Who typically uses this service
Dynamics 365 or Microsoft 365 is in place, alongside an older ERP, a shop, an accounting system, and a shipping solution. You want to connect the islands without setting up another mega-project — pragmatic, step by step, with clear ROI per interface.
Own shop, Amazon, multiple Shopify instances, plus B2B direct sales through Dynamics 365. The ERP should remain the single source of truth, with all channels hanging off it, inventory and orders flowing without Excel bridges.
SAP is in place for finance and production, with a Microsoft world coming on top for CRM, service, or field service. You need a reliable data flow between SAP and the Microsoft world — bidirectional, documented, audit-ready.
You're building your own product on Microsoft Power Platform or Dynamics 365 and need production interfaces to customer systems. We deliver the integration layer as sub-engineering — white-label, documented, handover-ready.
Related
Our own modular ERP with live integrations to Oxaion, Amazon, Shopify, and GLS — built on exactly the principles described on this service page.
Read more →The CRM and ERP platform we connect the majority of our interfaces to — with deep knowledge of Dataverse, plug-ins, and Power Automate.
Read more →Power Automate as the integrated flow layer, Dataverse as the data model, custom connectors for your own APIs — the platform layer underneath many of our interfaces.
Read more →Frequently asked questions
Live references with Oxaion ERP, SAP, Microsoft Business Central, Shopify (multi-shop), Amazon Reseller API, eBay, GLS, DHL, DPD, DATEV, lexoffice, Stripe, and PayPal. Plus Microsoft Graph for Microsoft 365 data and any REST, OData, or SOAP APIs. For unfamiliar platforms, we review the API situation in the initial conversation — the prerequisite is a documented endpoint with read and write permissions.
A cleanly scoped fixed-price interface (e.g., order push from Shopify into Dynamics 365 with status feedback) typically falls in the low five-figure range, depending on mapping complexity, error-handling requirements, and monitoring depth. More complex bidirectional master-data synchronizations or multi-tenant setups run higher. We estimate after a discovery call with a requirements document.
Either us in the Care model (monitoring, bug fixes, Release-Wave adjustments, ongoing development) or your team with handover documentation and knowledge-transfer sessions. The Care model is the rule for interfaces to third-party systems that regularly ship API changes themselves — e.g., Amazon, Shopify, Microsoft Graph.
We track Microsoft Release Wave announcements from the first public preview day and test all interfaces we built on the Wave preview environment before the production release rolls out. Anomalies are addressed directly in the Care model. As a result, an arades interface typically survives several Release Waves without a production outage.
Yes. SOAP, XML-RPC, classic web services, CSV/EDIFACT file drops over SFTP, direct database access, and even screen-scraping bridges for legacy host systems — we've built all of these in production projects. REST and OData are the norm today, but industrial and ERP legacy systems often bring older protocols with them — we either connect to them directly or wrap them behind a modern API facade.
A scoped fixed-price interface typically takes 4 to 8 weeks from kickoff — requirements workshop, mapping design, implementation, sandbox testing, monitoring setup, go-live. More complex bidirectional master-data synchronizations with conflict logic run 8 to 16 weeks. In ongoing sprint logic under time-and-materials, we ship a production-capable increment every 2 weeks.
Take-away · two materials
Two depths for different reading needs. The factsheet is a quick reference (3–5 min) and instantly downloadable. The whitepaper is market education with methodology and comparison data (15–30 min) — you get it by email after a short request.
3–5 min read · Direct download · no form
Compact overview: scope, key metrics, pricing model, process — ideal to forward to CFO, procurement, or the business unit.
15–30 min read · by email after request
Methodology, comparison data, recommendation framework — material for internal argumentation toward stakeholders.
Related services
When the integration requires deeper customization.
Read more →
When the integration is part of a larger project.
Read more →
Integration patterns, event models, data topology.
Read more →
Data migration during system changes — structured, with a defined cut-over strategy.
Read more →
30 min · discovery call
30-min discovery call with our integration engineering: we hear your requirement, check the third-party API situation, and come back with an effort estimate and a proposal for the delivery model that fits. You can see a concrete product example of our interface practice on the Intercompany product page.
