Platforms and integration

Business systems, joined so the operation can run.

This practice is about enterprise applications, SaaS platforms and the interfaces between them. Engineering is used to implement a platform decision and leave it comprehensible. It is one capability, beside architecture, configuration and integration.

How a platform is read

01

Experience

Sites, products and the work people actually do
02

Applications

SaaS platforms and business systems
03

Interfaces

APIs, events and the ownership of each record
04

Operations

Reporting, support and the cost of change

What a SaaS decision actually leaves you

Where the software is multi-tenant, the vendor runs the application infrastructure. The decisions that remain are tenancy, identity, configuration, integration and the cost of leaving. Where the organisation hosts the software, infrastructure is part of the same conversation, including cloud when it fits the operation. There is no standing preference beyond the two clouds that already hold most of this work: Microsoft Azure and Amazon Web Services. Kubernetes is used when the software is not a vendor's tenancy.

Interfaces, not a catalogue of APIs

An integration is a statement about who may call what, which system owns the record, and what happens when the receiver rejects the message. A modernisation case is the same kind of statement: which boundary stays, which one moves, and how the change will be estimated from those boundaries rather than promised as a fixed cost.

Connected applications are the normal case.

Few operations of any seriousness live inside one product. Finance, fulfilment, publishing, partner tracking and customer administration each arrive with their own system. The ones that recur are Microsoft Dynamics 365, Salesforce and SAP S/4HANA, with Microsoft Entra ID around them. The architecture is drawn with the C4 model and event storming, shaped with domain-driven design, and a live replacement is planned as a strangler fig rather than a cutover weekend. Interfaces run on Azure Service Bus, Amazon EventBridge, REST or GraphQL, whichever the two systems can sustain.

Related work sits with data and automation and digital strategy.