How Much Does Business Process Automation Cost for a UK SMB?
Business process automation for a UK SMB with 5-40 staff typically costs between £3,000 and £25,000 for a single well-defined project, with most engagements I run landing between £5,000 and £12,000. The final number depends on how many systems you're connecting, how messy your data is, and whether you need something built once and left alone or something maintained ongoing. Anyone quoting you a single fixed figure before seeing your systems is guessing. I'll walk through what actually drives the cost, show you the maths on payback, and tell you the two situations where I turn the work down outright.
How much does business process automation cost?
The honest answer is: it depends on scope, but here are the bands I see across engagements in Bristol and the South West.
- Single-process automation (one trigger, one or two systems, e.g. Xero invoice reminders or a form-to-CRM handoff): £2,500 to £6,000.
- Multi-step workflow (three or more systems, conditional logic, some data cleaning): £6,000 to £15,000.
- Departmental overhaul (finance, ops, or client onboarding rebuilt across four or more tools with reporting): £15,000 to £30,000+.
On top of build cost, budget 10-20% of the original build fee per year for maintenance, because APIs change, staff change, and business rules drift. I've had clients skip this and end up with a workflow silently failing for three weeks because a spreadsheet column got renamed.
What actually drives the price up or down?
Five things move the number more than anything else:
- Data quality. If your customer records in Sage or your property data in Reapit or Alto are inconsistent, someone has to clean that before automation works. I've spent more hours fixing duplicate supplier records than building the actual workflow.
- Number of systems involved. Two systems talking to each other is straightforward. Five systems with different authentication methods and rate limits is a different job entirely.
- Exception handling. The happy path is cheap. Building for what happens when a payment fails, a field is blank, or an API times out is where most of the hours go.
- Compliance and audit requirements. If you need a record of every action for accounting or regulatory reasons (common in property management and finance), that adds build time.
- Who maintains it afterwards. A one-off script nobody touches again is cheaper than a workflow your team will edit monthly, because the second needs documentation and a simpler interface.
What's a realistic worked example?
Let me show the arithmetic rather than assert a saving, because most automation pitches skip this part.
Say you employ an administrator on £26,000 gross salary, chasing rent arrears manually using data pulled from Reapit. Using a fully-loaded rate of 1.3x gross salary to account for employer's National Insurance, pension contributions, and overheads:
- Gross salary: £26,000
- Loaded factor: 1.3
- Loaded annual cost: £26,000 × 1.3 = £33,800
- Working hours per year (after leave, sick days): 1,650
- Fully-loaded hourly rate: £33,800 / 1,650 = £20.48
Now suppose that administrator spends 5 hours a week manually checking arrears reports and sending chase emails, and automation cuts that to 1 hour a week of review because the system flags overdue accounts and drafts the emails automatically.
- Hours saved per week: 4
- Working weeks per year: 46
- Total hours saved per year: 4 × 46 = 184
- Annual saving: 184 × £20.48 = £3,768
Against a build cost of £4,500 to £7,000 for that specific workflow, payback lands somewhere between 14 and 22 months. That's a reasonable case for doing it. If the same client wanted five different arrears workflows across five property types with different rules, the build cost would roughly double but the saving wouldn't, because the marginal admin time saved per extra workflow shrinks. That's the calculation I ask every client to do before signing anything.
Should I build this myself with Zapier, Make or n8n instead?
You can, and for simple single-trigger workflows I'd tell you to. If you need to get a Google Form submission into Xero as a draft invoice, or push a new lead from a website form into a CRM, a no-code tool will get you there in an afternoon for the cost of a subscription, typically £15 to £70 a month depending on volume.
Where these tools start costing you more than they save is volume and complexity. Per-task pricing on Zapier and Make gets expensive quickly once you're processing thousands of records a month. n8n is cheaper to run if you're comfortable self-hosting, but then you've taken on server maintenance, which is its own job. I've taken over three client accounts this year where the original no-code workflow cost more in subscription fees than a properly built, purpose-specific script would have cost to run. The trade-off isn't good versus bad, it's speed-to-build versus cost-at-scale. Know which one matters more for your situation before you commit.
When would you turn this work down?
Two situations, both of which I've walked away from in the last year.
First: when the underlying process is broken, not slow. If a client wants me to automate a workflow that exists because of an organisational disagreement about who owns a decision, automating it just makes the argument happen faster and with less human judgement in the loop. I had a client want automated approval routing for purchase orders where the real problem was that two directors disagreed on spending limits. No script fixes that. I told them to sort the policy first and come back.
Second: when the volume doesn't justify it. If a task takes someone 20 minutes a week, and building the automation would take 15 hours at a fully-loaded consulting rate of £75 to £120 an hour, you're looking at a payback period measured in years, if ever. I turned down a hospitality client last spring who wanted booking confirmations automated across three systems for a task that took their receptionist 15 minutes a day. The maths didn't work, and I said so.
If a consultant never turns down work on these grounds, they're optimising for their invoice, not your outcome.
What ongoing costs should I budget for after launch?
Budget 10-20% of build cost per year for maintenance, plus whatever subscription fees the underlying tools carry. Xero, Sage, and Companies House integrations are generally stable because they're widely used and well documented. Smaller or niche software, including some property and legal case management systems, changes its API with less notice, and that's where maintenance hours actually go. I put a maintenance clause in every contract now specifying response time and monthly review, because "it broke and nobody noticed for a month" is the single most common failure mode I see in automation projects built by someone who then disappeared.
Frequently asked questions
Is automation cheaper than hiring another admin person?
Often yes for repetitive, well-defined tasks, but not always. Compare the annual saving in hours (as shown above) against build plus maintenance cost over three years. If the task changes frequently or needs judgement, a person is usually the better investment.
How long does a typical project take to build?
A single-process automation usually takes one to three weeks from scoping to live. Multi-step workflows with several systems typically take four to eight weeks, mostly because of data cleaning and testing exception cases, not the actual build.
Do I need to replace my existing software to automate?
Almost never. Most of the work I do connects existing systems, such as Xero, your CRM, and your booking software, rather than replacing any of them. Replacing software is a much bigger and separate decision.
Can I automate something and then take it in-house later?
Yes, and I build for that deliberately. I document workflows in plain language, not just code comments, specifically so a client's own staff can take over maintenance once they're confident, if that's what they want.
Where to start
Before you spend anything, pick one recurring task that happens at least weekly, time it honestly for a week using a stopwatch or your phone timer, and run the arithmetic above using your own salary figures and a 1.25 to 1.4x loading factor. If the annual saving is under £1,500, don't automate it yet. If it's over £3,000, you've got a solid business case to bring to whoever you talk to next, whether that's me or someone else.
If you want a second opinion on whether a specific process is worth automating before you spend a penny, I offer a 20-minute call to talk through the numbers.