Workflow automation
The admin that eats your evenings, handled.
Most small businesses lose more money to unanswered enquiries and unchased quotes than to anything on their website.
Werehouse Media builds workflow automation for UK small businesses: enquiry routing, quote follow-up, review requests, invoice chasing, booking reminders and monthly reporting. It is priced on application rather than from a fixed number, and that is deliberate. The same job genuinely costs different amounts in different businesses: chasing quotes for a firm that already keeps them in one system is a morning, and the identical brief for a firm holding them across email, a spreadsheet and a notepad is several days of untangling before any automation gets built. Quoting a headline price before seeing which of those you are would mean either overcharging the tidy business or discovering the mess halfway through and asking for more. Every automation is quoted in writing after a short call, fixed before work starts, and built on tools you own and keep. Most projects pay for themselves in recovered admin time rather than in new customers, so the honest test is how many hours a week it gives you back.
Pricing
Why is automation priced on application?
Because the same brief is a different job in two different businesses, and any honest quote has to account for that. Everything else on this site carries a published price. This one does not, and here is exactly why.
Take quote chasing. In a business that already keeps every quote in one system with a date against it, the automation is a morning's work: read the list, wait the agreed number of days, send a polite follow-up, stop when the quote is accepted. In a business holding quotes across an inbox, a spreadsheet and a notepad in the van, the same brief means untangling where the data actually lives before a single automation can be built. That is several days, and it is the more common case.
A headline price would mean one of two dishonest outcomes. Either the tidy business subsidises the messy one, or the messy one gets quoted low and then asked for more money halfway through, which is the thing everyone hates about this industry.
What decides the price?
| Factor | Cheap end | Expensive end |
|---|---|---|
| Where your data lives | One system, consistently used | Spread across inbox, spreadsheets and paper, with no shared reference |
| How many exceptions | One path, handled the same way every time | Rules that change by customer, product or season |
| What it connects to | Tools with a proper API and a documented integration | Legacy software that only exports a CSV, or nothing at all |
| Who has to approve | Sends automatically once set | Needs a human check before anything reaches a customer |
| What happens when it fails | A retry and an alert to you | Regulated or financial steps needing an audit trail |
You get a fixed price in writing after a short call, before any work starts. No hourly meter and no surprise invoice, same as every other service here.
What can actually be automated?
- Enquiry routing and acknowledgement. Every enquiry answered within seconds, tagged, and put somewhere you will see it. The single highest-return automation for most trades, because the business that replies first usually wins the job.
- Quote follow-up. A polite chase on a schedule you set, stopping the moment a quote is accepted or declined.
- Review requests. Sent after a job completes, to every customer rather than only the ones you expect to be happy, which is also what the reviews rules require.
- Booking and appointment reminders. Fewer no-shows, no calls from you.
- Invoice chasing. The conversation nobody wants to have, had automatically and politely.
- Monthly reporting. Enquiries, sources and outcomes assembled and sent to you, rather than you building it.
What should not be automated?
Worth saying, because the answer is often more than people expect. Anything where a customer needs to feel a person made a decision. Complaints. Quotes for unusual work. Anything a regulator expects a named human to sign. Automating those saves an hour and costs a relationship.
I will also say no where the volume does not justify it. If you send four quotes a month, chase them yourself. The automation costs more than the time it saves, and I would rather tell you that on the call than take the money.
What do you end up owning?
The automations run on your accounts, under your logins, on tools you pay for directly and can cancel. Nothing is locked to me, and if you stop working with me it keeps running. You get written documentation of what fires when, so anyone can pick it up. Same principle as the websites: you own it outright.
Most of this pays for itself in recovered admin time rather than in new customers, so the honest measure is how many hours a week you get back. That is the number I will ask you to judge it on.
Straight answers
Common questions.
Because I would have to pick a number that is wrong for most readers. Every other service here is published precisely because a website build is predictable enough to price in advance. Automation is not: the work is mostly determined by the state of your existing systems, which I cannot see from here. A short call tells me, and then you get a fixed figure in writing.
Mostly ordinary automation, and that is the right answer more often than the industry admits. Rules, triggers and integrations are predictable, cheap to run and easy to debug at three in the morning. AI gets used where genuine judgement is needed, such as summarising a long enquiry. Putting a language model in charge of sending your customers invoices is a way of making a reliable process unreliable.
Every automation is built to fail loudly rather than silently. If something cannot complete, it retries and then tells you, rather than quietly dropping an enquiry. Silent failure is the real risk with this work, which is why alerting is part of the build rather than an extra.
Next step
What takes up your week?
Tell me the job you keep doing by hand. I will tell you whether it is worth automating, including when the answer is no.