Integration Platforms
iPaaS, middleware or custom APIs: pick the integration platform that fits your data, logic and compliance needs. Talk to us.
Quick answer:An integration platform connects systems like Xero, HubSpot and Shopify. The right type (iPaaS, middleware or custom APIs) depends on data volume, logic and compliance needs, not business size.
- Platform Engineering
- System Integration
- API Development
- Cloud Engineering
Quick answer
What is an integration platform, and when does a business actually need one?
An integration platform connects systems such as Xero, HubSpot and Shopify so data moves automatically. The right type (iPaaS, middleware or custom APIs) depends on data volume, logic and compliance needs, not company size.
Additional Context
Choosing between the three is a connection-by-connection decision, not a single platform purchase.
Sources
- Consumer Data Right
Explains the standardised API requirements the Consumer Data Right places on banking, energy and telecommunications providers.
- Australian Government API Design Standard
Sets mandatory design conventions for APIs built by Australian Government agencies, widely referenced outside government too.
Platform Engineering
What an integration platform actually does
An integration platform is the layer of software that moves data between the systems a business already runs. For a business running Xero or MYOB for finance, HubSpot for sales and Shopify for orders, it is whatever mechanism keeps an invoice, a customer record or a stock level consistent across all three, without someone copying numbers between browser tabs. That mechanism might be a commercial iPaaS tool, a middleware layer sitting between systems, or a set of purpose-built APIs. The term covers all three because the problem they solve is the same: the same data needs to exist correctly in more than one place.
iPaaS, middleware or custom-built: which layer fits
Off-the-shelf iPaaS tools such as Zapier, Make or Workato handle straightforward, low-volume connections well: a new Shopify order creating a Xero invoice, a HubSpot form filling a spreadsheet. They become expensive and brittle once the logic gets conditional, the volume climbs into thousands of records a day, or the connector for a particular system doesn't exist. At that point, teams typically move to middleware that centralises transformation logic, or to contract-first API design built specifically for the systems involved. The decision is rarely one-size-fits-all across a business: a company might run iPaaS for marketing automation while building custom APIs for the finance-to-warehouse pipeline that touches every order. System integration work is usually a mix of all three, chosen connection by connection.
Where manual re-entry turns into risk
Integration Platforms
Problem
Finance, sales and operations teams manually re-key the same data across Xero or MYOB, HubSpot and Shopify because no reliable layer keeps those systems in sync, and every manual step is a point where numbers drift apart.
Common Pain Points:
- Duplicate data entry The same invoice, order or customer record is typed into more than one system, and the versions drift apart within weeks.
- No one owns the sync logic A script or workflow built years ago still runs, but nobody currently on staff knows what it does or when it last worked correctly.
- Scaling breaks the shortcut A no-code workflow that worked at low volume starts timing out, dropping records or hitting rate limits as order volume grows.
Business Impact:
Time Wasted: Staff time lost to manual re-entry and reconciliation checks
Cost Implication: Recurring cost of duplicate work, licensing overlap and error correction
Opportunity Cost: Finance and operations staff spend cycles reconciling spreadsheets instead of improving service delivery or forecasting
Solution
National Digital assesses which connections need a commercial iPaaS tool, which need middleware, and which warrant purpose-built APIs, then builds only what the data volume and logic actually require.
Our Approach:
- Map the data and decision points
Identify every system touching the same record, where business logic branches, and where compliance or audit requirements apply.
- Build and stage the integration layer
Implement the chosen combination of iPaaS, middleware or API work against a staging copy of production data before any live cutover.
Expected Outcome: A single source of truth for shared records, with manual re-entry replaced by monitored, documented data flows between systems.
Key Takeaways
What matters when choosing an integration platform
- The right integration layer is decided by data volume and logic, not company sizeImportant
A small business with complex conditional logic may need custom APIs sooner than a larger one running simple, high-volume connections.
- iPaaS tools are a genuine fit for straightforward, low-volume connectionsImportant
Tools like Zapier or Make handle simple triggers well and remain the right choice when the logic stays simple and volume stays modest.
- Ownership of integration logic needs to sit somewhere specificImportant
Integrations built as one-off scripts fail quietly when the person who built them leaves, so documented ownership matters as much as the build.
- Regulated data flows change the calculus for build-versus-buy decisionsImportant
Where personal or financial data moves between systems, the business carries obligations under the Privacy Act regardless of which tool moves the data.
Choosing an integration platform means matching the tool to the data volume, business logic and compliance load actually present, not defaulting to the most popular no-code option.
Why API standards and data rules matter here
Integration platforms increasingly move regulated data between systems, which is why government API standards and privacy obligations shape architecture decisions, not just technical preference.
Banking, energy, telco
Consumer Data Right
Significance: highThe Consumer Data Right requires banks and other designated sectors to expose standardised APIs for sharing customer data on request, a model private integration projects increasingly mirror.
Source: accc.gov.au – Consumer Data Right
Mandatory for government APIs
API Design Standard
Significance: mediumThe Digital Transformation Agency's API Design Standard sets mandatory conventions for APIs built across Australian Government agencies, widely referenced by private-sector integration teams.
Source: api.gov.au – API Design Standard
Applies to all APP entities
Notifiable Data Breaches scheme
Significance: mediumAny organisation covered by the Australian Privacy Principles must notify eligible data breaches, a direct consideration whenever personal data moves between integrated systems.
Source: oaic.gov.au – Notifiable Data Breaches scheme
Methodology
Drawn from publicly available regulatory guidance published by the ACCC, the Digital Transformation Agency and the OAIC, current as of the review date above.
Spotting the failure pattern early
Signs your current integration layer is already failing
The pattern shows up the same way across most finance and operations teams: someone exports a report from one system, reformats it in a spreadsheet, and re-enters it into another. It is rarely flagged as an integration problem; it is flagged as the month-end close running long, or stock discrepancies nobody can explain. A recurring failure mode in ERP data synchronisation projects is discovering that three different systems each hold a slightly different version of the customer master file, none of them authoritative, because nobody owns the sync logic between them. The same thing happens on the sales side: a CRM to accounting system sync built years ago as a one-off script quietly stops updating a field, and nobody notices until a customer disputes an invoice total that doesn't match the quote the sales team sent.
Getting the architecture right for the scale you're at
The right answer depends on how much the integration actually has to carry, not how large the business looks on paper. A retailer pushing a few hundred orders a day through Shopify into Xero has a genuinely different problem from a logistics operator synchronising inventory across four warehouse systems in near real time. Building custom infrastructure for the first case is usually over-engineering; relying on a brittle no-code workflow for the second is how outages happen during a peak trading period. The practical test is whether the integration needs to handle conditional business logic, whether it touches regulated data, and whether a failure would be visible to customers within hours rather than caught at month-end reconciliation. Those three questions, more than headcount or revenue, determine whether an iPaaS tool is enough or whether the business needs a properly engineered integration layer.
Integration platform questions, answered directly
What's the difference between an integration platform and an iPaaS tool?
An integration platform is the broader category: any combination of tools and code that keeps systems in sync. iPaaS (integration platform as a service) is one option within that category, built for low-code, trigger-based connections. Middleware and custom APIs are the other two, used when logic or volume outgrows what iPaaS handles well.
Is platform engineering the same thing as DevOps?
They overlap but are not identical. DevOps describes practices for shipping software reliably. Platform engineering is the discipline of building the internal tools, APIs and infrastructure that make those practices usable day to day, typically serving the organisation's own development and operations teams.
What does an API actually do when it connects two business systems?
An API defines a fixed way for one system to ask another for data or trigger an action, for example Shopify asking Xero to create an invoice when an order settles. It replaces manual export-and-import steps with a direct, repeatable call that runs the same way every time, which is why well-designed APIs cut the error rate that manual re-entry introduces.
When does a business outgrow tools like Zapier or Make?
Usually when the logic stops being a simple trigger-then-action chain. Conditional branching, high record volumes, strict error handling or regulated data all push a business toward middleware or custom-built APIs, because no-code tools are priced and designed for straightforward connections, not complex business rules.
What's the actual benefit of investing in proper system integration?
The direct benefit is that a number entered once stays correct everywhere it appears, so finance isn't reconciling differences between systems at month-end and sales isn't quoting a price that doesn't match what's in the accounting system. The indirect benefit is capacity: staff previously doing manual re-entry can be redeployed to work that needs judgement.
How is platform engineering different from regular software engineering?
Software engineering builds the product customers or staff use directly. Platform engineering builds the layer underneath it: the APIs, pipelines and infrastructure that other engineering teams rely on to ship their own work. For a growing business, platform engineering usually means the team responsible for integration and automation, built into existing IT operations.