--- title: "The under-£1,000 automation: what's actually possible" description: "Under £1,000 buys certainty, not software: an Audit from £450 sizes the job. One well-scoped workflow fits a Workflow Sprint, from £1,950." url: "https://automancer.uk/field-notes/under-1000-automation-whats-actually-possible/" canonical: "https://automancer.uk/field-notes/under-1000-automation-whats-actually-possible/" type: "BlogPosting" author: "Waseem Ilyas" datePublished: "2026-10-05T00:00:00.000Z" --- Say "automation project" to most small-business owners and they brace for five figures and six months. Sometimes that's the right size. Quite often, though, the job is one annoying workflow, and the honest price for one annoying workflow is a lot smaller than the fear. It's just not quite as small as the internet would like you to believe. ## How much does a small automation project cost? With us, a small automation project starts at £1,950. That's a [Workflow Sprint](/services/): 1–3 automations live in around two weeks. Under £1,000, what you can reliably buy is certainty rather than software. An Automation Opportunity Audit, from £450, maps where your time is going and hands you a costed plan for fixing it. We publish those prices so you can do this sum before talking to anyone. Every price on our services page is a *from* price for real work, not a teaser rate. If anyone offers to automate your whole business for a few hundred pounds, ask what happens the first busy Monday it breaks. ## What a small job looks like: one workflow, start to finish Take leave requests. Written down, it sounds trivial: someone asks for time off, someone else says yes. In a business with dozens or hundreds of staff, it's forms going missing, approvals sitting in whoever's inbox they landed in, a calendar nobody fully trusts, and no record to point at when two people remember a conversation differently. A well-scoped version does four things: - **Request.** Staff ask for leave in one place, not on a scrap of paper. - **Routed approval.** The request goes to the person who should approve it, not whoever happens to be nearest. - **Calendar.** Approved leave lands on a calendar view the rota can be planned against. - **Audit trail.** Every request and decision is recorded, so "I asked in March" can be checked rather than argued about. That's the whole thing. No new HR system, no payroll integration bolted on because it seemed a shame not to. Approval routing is one of the jobs we list as a good fit for a Sprint, alongside enquiry-to-quote handling and the report that currently eats someone's Friday. It's not a niche problem, either. Support-worker leave requests are one of the pieces in active build for [a ~600-person supported-living care provider](/work/care-provider-transformation/), as one piece of a programme that runs to roughly eighteen governed workstreams. Small pieces sit perfectly happily inside big programmes. They just have to be scoped as small pieces. ## Small and cheap is a scoping decision, not a corner cut A small job isn't small because anything was skipped. It's small because the scope was disciplined from the first conversation. There's no trick to it: it's scope discipline, which is less exciting than magic and considerably more reliable. Three things keep a job small: - **One workflow, not a platform.** Request, route, approve, record. Everything adjacent is left out on purpose, and written down as left out. - **Rules that can be written down.** If you can say who approves what, the build is mostly plumbing. If that's contested internally, settling it is the expensive part, and it has to happen before anyone builds. - **A place it already lives.** A workflow that plugs into the tools you use every day stays small. One that needs a new system to sit in grows. Some things genuinely can't be this small. The same care provider's core platform turned messy CSVs into 294 clean carer records and auto-created 588 supervision schedules. That's a System / Agent Build, from £4,500, because there is more engineering in it. The point isn't that everything is cheap. It's that a small problem deserves a small price, not one inflated because "automation" sounds expensive. ## How to spot your own small candidate Before you ask anyone to quote, put your candidate through four questions: 1. **Can you describe it as "when X happens, do Y"?** If the whole job fits in that sentence, it's small. If it needs its own users, permissions and data model, it isn't, and that's fine. It's just a different kind of project. 2. **Do you already know who decides what?** If the approval logic is clear, the build is mostly plumbing. If it's disputed, sort that out first, because no software settles an argument your team hasn't had yet. 3. **Can it live inside something you already have?** If yes, it stays small. If it needs a new home, budget for building the home too. 4. **What does the manual version actually cost you?** Count the hours and the mistakes honestly. If the answer is "not much", leave it alone, however satisfying it would be to fix. If the first three answers are yes and the fourth is "more than we'd like to admit", you probably have a small job on your hands, not a programme. ## Where under £1,000 goes furthest If you can't tell whether your problem is small or big, spend the first £450 finding out. The [Audit](/services/) gives you a ranked list of what to automate first, with an estimate of what each item costs you now. It's your plan to keep, and if you go on to do a Sprint or a Build with us, the fee comes straight off the price. If you already know the job and it passes the four questions above, a Workflow Sprint is the right shape for it, and it has room for a couple more like it. And if you already have automation live and want it watched and extended, our AI Ops Partner retainer is from £495/mo, cancel anytime. Either way, a person reads what you send and writes back. No funnel, no countdown, just a sensible conversation about the size of the job.