An integration is just a connection that lets two of your software tools share data automatically, so you stop copying it by hand. When a customer pays in your shop and the sale appears in your accounting tool without you touching anything — that's an integration at work. Most small businesses need a connector tool like Zapier or Make, not custom code, to achieve this. The short answer: before you pay a developer, check whether your tools have a built-in connection or whether a no-code connector can handle it.
If you're currently logging into three different systems to enter the same order, the problem has a straightforward solution. You don't need a bespoke software project — you probably need to spend an afternoon setting up a connector that already exists. This post explains the options, in the order you should try them.
Why disconnected tools cost you money
When your tools don't talk to each other, the work falls on you. You type the same order into your shop, your shipping tool, and your accounts. You export a spreadsheet from one system and import it into another every Monday morning. You catch a discrepancy at month-end because two systems counted the same transaction differently.
The hidden cost of manual data entry isn't just the time spent doing it — it's the errors that creep in and the delays in acting on information. An order that takes two minutes to copy becomes an order that was copied wrong and costs an hour to untangle. A weekly import that runs fine ninety-nine times fails on the hundredth and corrupts a month of records. The compounding cost of disconnected tools is usually higher than people expect before they start measuring it.
The three ways to connect tools
There's a hierarchy here, and working through it in order saves money.
Native built-in integrations — free, try first
Many popular tools have built-in connections to other popular tools. Shopify connects directly to Klaviyo. QuickBooks connects directly to Stripe. Google Calendar syncs with most scheduling tools without any third-party software. Before you look at anything else, check the integrations or app marketplace inside each tool you're already paying for.
Native integrations are free, maintained by the tool vendors, and usually the simplest to set up. They're also the most reliable because the vendor updates them when either tool changes. The limitation is that they only cover popular combinations — if your stack includes anything niche or old, the native integration probably doesn't exist.
No-code connectors — subscription, no developer
If there's no native integration, tools like Zapier, Make, or n8n let you build connections between thousands of apps without writing any code. You define a trigger (a new form submission, a new order, a new contact) and an action (create a row in a spreadsheet, send a Slack message, add the contact to an email list). The connector handles the data transfer.
This is the right tool for most SMB integration needs. The subscription fee — typically in the $50–200 per month range you'll see quoted — is usually a fraction of what a developer would charge to build the equivalent, and you can usually set it up yourself in an afternoon. Connectors handle an enormous range of use cases, and their marketplaces have pre-built templates for many common workflows.
I'd encourage you to try a connector before calling a developer for almost every integration need. Most of the time, it's enough. This is the same honest advice I give in the comparison of off-the-shelf AI tools versus custom automation — custom work earns its place only once you've confirmed the simpler option won't reach.
Custom API integration — for what connectors can't do
When a connector won't work, a developer builds a direct integration against the tool's API — a programming interface that lets software talk to software. This is more expensive and takes longer, but it handles cases the connector can't.
A custom integration is the right call for complex logic (multiple conditions, data transformation, error handling), old software that no connector supports, or high task volumes where per-task connector pricing adds up faster than a one-time build. It also makes sense when data needs significant reshaping before it moves — pulling a field from one system and splitting it into three fields in another, for example.
When a no-code connector is enough
Two popular, actively maintained tools with a connector between them: use the connector, don't pay a developer. If your workflow is a simple trigger-and-action — new contact in your CRM starts an email sequence, new paid order updates your inventory spreadsheet — a connector handles this reliably and for a monthly fee that's well inside what most businesses already spend on software.
The honest threshold: if you can describe your integration as "when X happens in tool A, do Y in tool B," a connector almost certainly handles it. Test the free tier first; if the automation runs without hitting limits, you've found your answer cheaply.
When you need custom work
Some situations genuinely need a developer:
Old software with no modern API. Accounting systems from fifteen years ago, bespoke industry tools, anything that was never designed to connect to the outside world. These often can't be reached by a connector at all. A developer can sometimes work around this by processing file exports or reading a database directly, but it requires custom work.
Complex transformation logic. If the data needs significant reshaping before it moves — combining fields, applying business rules, handling exceptions differently based on customer type — a connector's visual workflow builder becomes painful or impossible to maintain. Code handles this more cleanly.
High task volumes. No-code connector pricing often scales with the number of tasks (individual data transfers) per month. At low volume this is inexpensive. At high volume, the monthly cost can exceed what a one-time custom build would have cost within a year. If you're running tens of thousands of automated tasks per month, do the maths before renewing your connector subscription.
Two-way syncing between systems. One-way flows are straightforward. Keeping two systems genuinely in sync — where changes in either one update the other — involves conflict resolution logic that connectors handle poorly. This is usually custom work.
Where AI fits in
Most integrations are mechanical: copy a value from here to there. AI adds something useful when the data needs to be understood before it moves — when raw input isn't yet in a shape the next tool can act on.
A customer support email arrives. Before it goes into your helpdesk, AI classifies it: is this a complaint, a refund request, a general question, or a sales enquiry? The classification then routes the email to the right queue or assigns the right priority. The integration still moves the email; the AI layer adds the decision.
Similarly, AI can extract structured data from unstructured inputs. An emailed invoice with a free-text layout, a PDF order form, a voicemail transcription — AI can pull out the relevant fields (supplier name, line items, amounts, due date) and pass clean, structured data into your accounting or inventory tool, without anyone retyping it. The same applies to classifying incoming messages by intent as they arrive in a shared inbox, letting the integration route them automatically rather than relying on a human triage step.
The AI cost at small-business volumes is metered and usually modest — worth factoring in, but rarely a deciding factor. What matters more is getting the underlying integration solid first. Add AI once the basic data flow is working reliably; it is a layer on top, not a starting point.
What it costs
A no-code connector subscription runs $50–200 per month in typical ranges. A custom integration project — scoped, built, and tested — typically falls in the $2,000–$10,000 range for a focused piece of work. Larger projects that connect multiple systems or include complex business logic scale from there.
If you want the full breakdown of what automation work costs at different levels of complexity, the business automation cost post covers that in detail. If you're not sure which process to tackle first, the what to automate first post gives a practical framework for prioritising.
For most small businesses, the honest starting point is: check for a native integration (free), then try a connector (subscription), then call a developer only if neither reaches. The majority of integration problems are solved before step three.
If you're looking at a specific integration and want a straight answer on whether custom work is warranted, take a look at what I offer — I'm usually able to give a clear direction after a short conversation.
