API Integration Services: What They Actually Do and Cost

A construction management platform I came across will not let you overwrite a cost catalogue. To update it, you have to delete every existing line first, one at a time, because the import creates duplicates rather than replacing what is there. There is no bulk delete button.

In practice, people export the whole catalogue, change it, delete everything by hand, and reimport it fresh.

That is not a technology problem AI fixes and it is not a case for an integration either. Sometimes the honest answer is that the software itself is the problem, built awkwardly some years ago and never revisited. A consultant who tells you that is worth more than one who quotes you a project anyway.

What an API integration actually is

An API is the bit of a piece of software that lets other software talk to it. Xero has one. Sage has one. An integration uses that door to move data automatically, instead of someone exporting a spreadsheet every Friday and emailing it to the accountant.

A sale in your till system can appear in Xero as an invoice without anyone typing it in. A new employee in your HR system can create a payroll record without their tax code being entered twice.

That is different again from Zapier style connectors, which usually still work through each system's API, but via a shared platform that configures the connection for you rather than someone building it from scratch. For most small businesses that configured route is cheaper, faster, and entirely sufficient. If you want the wider picture of where that kind of connector fits against a proper rebuild, we have written about that split before.

System integration services, in the fuller sense, covers all of it: a quick app setup, a bespoke build, and the judgement call about which one a given job actually needs.

When you need one, and when you do not

You need an integration when the same information is being typed into more than one system, regularly, by a person. Supplier invoices into the accounts package and also into the job management system. Timesheets into payroll and then again into a costing spreadsheet.

You do not need one if the task happens once a month and takes ten minutes. The cost of building and maintaining an integration can easily outweigh the time it saves, and that is worth saying plainly because it is not the advice you get from an agency that sells integration work. If duplicated data entry is only part of a wider mess of spreadsheets and manual checks, a process map usually surfaces it faster than a tool does.

Sage integration

Sage integration usually comes up for firms who have had Sage for years, often Sage 50, and are now trying to connect it to something newer. A CRM, perhaps, or a stock system.

Not every Sage product offers the same route in, and the difference matters more than people expect.

Sage Accounting, the cloud product, has a modern API, its own app marketplace and a Zapier connection. Linking it to something else is usually a configuration job rather than a build.

Sage 50 in the UK is a different thing. It has no modern web API. Third party software connects to it either through Sage's paid developer kit or through a read only link to the database. That is why an older Sage 50 setup is a longer job than it looks from the outside, and why a quote for one tells you very little about the other.

If what you actually need is help getting Sage, Xero or a booking system talking to each other properly, that is the specific job our accounting integration work covers. This post is the wider picture of how API work generally gets scoped and priced.

Xero integration apps, and where SumUp fits

Xero has built itself to be a hub that other software plugs into, and its app marketplace is one of the larger ones in UK accounting software. Payroll providers, point of sale systems and card readers mostly connect to it through a published app rather than a bespoke build.

SumUp Xero integration is a common request for firms taking card payments on the move. SumUp connects to Xero directly, through its own app in the Xero App Store. It sends daily card sales into Xero as invoices, split by VAT rate or by product, so nobody is keying card takings in alongside the till receipts.

Payroll software that integrates with Xero is the other frequent ask. Several payroll providers connect directly, so pay run totals reach the accounts without gross pay and the deductions being re-keyed by hand each month. Which one fits depends on your payroll provider and how your PAYE scheme is set up, which is worth working out before you pick either piece of software, not after.

Off the shelf is usually the right starting point

For a lot of small businesses, a published app in the Xero or Sage marketplace, set up properly with the mapping checked, covers what they need. A custom built API connection is for the genuinely unusual job, where nobody has built the link before.

I would rather spend an hour with a client working out whether an existing app does the job than quote them for a build they did not need. That costs me the build fee. It is still the right order to do things in.

What it costs, and what moves the number

There is no single price for this, because two different jobs hide under the same phrase. Configuring an existing marketplace app, with the mapping checked properly, is a bounded piece of work. Building a direct connection between two systems that were never designed to talk to each other is a different job entirely, and it gets priced as its own project once someone has actually looked at both systems.

A single connection starts from £500. Anything bigger than that is quoted after someone has looked at both systems, because the size of the job is not visible from the outside.

What changes the number in practice: how many systems are involved, how clean the underlying data is, and whether the integration needs to run both ways or just one. A firm with one messy, customised Sage 50 setup is a longer conversation than a firm plugging a new card reader into standard Xero.

The honest version is that nobody can give you a figure for the bigger jobs before looking at your systems, and anyone who does is guessing. What a fair consultant can do is scope it against what you actually have, then price it as one project rather than a large single number that is easy to defer.

Where this is not for you

If your business runs one system, your bookkeeper does the entry once a week, and nobody is retyping anything twice, you do not need an integration. You need nothing. A broader digital transformation review is the wrong next step too, if the only thing missing is a report nobody has asked for.

And if what you are actually after is the figures landing in front of you every Monday, that is an automated report rather than an integration. It is a smaller piece of work and a cheaper one.

Most businesses asking about this do not actually need a bespoke build. They need an hour spent working out whether a published app already does the job.