Process automation

Stop paying senior people to copy-paste.

The average mid-sized team loses 15–20% of operational hours to work that already lives inside the SaaS tools they're paying for — it just isn't connected. We connect it.

The problem

Your tools work. The handoffs between them don't.

Every growing business hits the same wall. The CRM is fine. The accounting tool is fine. The project management tool is fine. But invoices arrive in email, get downloaded by hand, renamed, uploaded somewhere, then keyed into accounting. New clients fill out a form, then someone manually creates the Slack channel, the Drive folder, the contract, and the kickoff invite. Approvals happen in DMs that no one can audit.

Each of these handoffs costs five minutes. Multiply by ten people, ten times a day, fifty weeks a year, and you're paying for a full-time job you didn't realise you'd hired. Worse, the work is invisible — until someone leaves and the institutional knowledge leaves with them.

Process automation isn't about replacing people. It's about ending the conditions where senior people quietly drown in clerical work.

Approach

How we solve it

Four phases. Two weeks for the first workflow in production. Then we keep going if it's working — and stop if it's not.

  1. 1

    Audit before automating

    We map the process end-to-end across people, tools, and inputs. Half the wins come from killing steps, not connecting them — we'll show you which ones to delete before we wire anything up.

  2. 2

    Pick the right tier of tool

    Zapier when speed matters and the workflow is short. Make for branching logic and bigger data. n8n (self-hosted) when you need control, scale, or compliance. We'll tell you which — and why — before we build.

  3. 3

    Build with real error handling

    Every workflow ships with retries, idempotency keys where it matters, alerting to Slack or email when something fails, and a runbook your team can follow without us.

  4. 4

    Hand it over, properly

    Credentials in your name. Documentation in your wiki. A 30-minute Loom walking your team through what to do when something breaks. You can fire us tomorrow and nothing stops working.

Stack

Tools we work in

We're tool-agnostic. We pick what fits your team's existing stack, your data sensitivity, and your budget — and we'll happily push back on a brief that demands a specific platform when something else is the better fit.

  • Zapier
    speed & breadth
  • Make
    branching logic
  • n8n
    self-hosted
  • Power Automate
    Microsoft stack
  • Airtable
    ops database
  • Notion
    docs + workflows
  • Google Workspace
    Apps Script
  • Slack
    approvals + alerts
  • Stripe
    billing flows
  • Xero / QuickBooks
    accounting
  • DocuSign
    signature workflows
  • Custom code
    when it earns it
Real-world

What this looks like in practice

Figures are typical ranges from comparable engagements

  • 11.5h
    Saved per week

    Invoice processing for a UK accountancy

    A 22-person practice was spending 12 hours per week downloading client invoices, renaming files, and entering them into Xero by hand.

    We built a Make scenario that pulls invoices from inboxes and shared drives, OCRs them, matches the supplier, and pushes draft bills into Xero with the right tax codes. The bookkeeper now spends 30 minutes per week reviewing exceptions.

  • 3 days → same day
    Onboarding cycle time

    Client onboarding for a US marketing agency

    A 40-person agency was losing the first three days of every engagement to a sequence of forms, contracts, intro emails, Slack invites, and folder creation handled by an over-stretched ops manager.

    A single intake form in HubSpot now triggers a Zapier sequence that fires the contract via PandaDoc, provisions the Slack channel, creates the Google Drive folder structure, schedules the kickoff in Calendly, and posts the welcome message — all before the new client refreshes their email.

  • ~$80k/yr
    Operational cost saved

    Approval routing for a 200-person SaaS

    Procurement requests bounced between three approvers in email, with no audit trail, no SLA, and a heroic ops lead chasing approvals on Slack.

    We replaced the email chain with an n8n workflow on top of an Airtable approvals table, with role-based routing, automatic reminders after 48 hours, and a single dashboard the CFO actually checks. Approval times dropped from a median of 9 days to 1.5 days.

Frequently asked questions

What's the difference between Zapier, Make, and n8n?
Zapier is the fastest to ship and has the widest catalogue of integrations — best when the workflow is short and you'd rather not run infrastructure. Make (formerly Integromat) gives you visual branching, error routes, and much cheaper pricing at higher volumes — best for anything with conditional logic. n8n is self-hostable and open source — best when you need full control over your data, want to avoid per-task pricing, or have compliance requirements that rule out third-party platforms. We'll recommend whichever is right for your stack and budget; we have no financial relationship with any of them.
Will this break when our SaaS apps update their APIs?
Sometimes — APIs do change. That's why every workflow we ship includes monitoring, automatic failure alerts, and a documented runbook. If you're on a retainer, we cover the fix. If you're not, you can fix most issues yourself in under 30 minutes using the runbook, or send us a one-off ticket.
Who owns the credentials and accounts?
You do. Always. We connect to your Zapier / Make / n8n account using your SaaS credentials. When the project ends, we hand back full admin access and remove ours. No vendor lock-in by accident.
How do you handle errors and edge cases?
Every workflow has error handling baked in: retries with exponential backoff for transient failures, dead-letter handling for malformed inputs, idempotency keys on operations that shouldn't run twice, and Slack/email alerts when something needs human attention. We also write a one-page "what to do when X breaks" runbook for each workflow.
Do we need engineering involvement on our side?
Usually no — that's the whole point. Most of the tools we use are no-code or low-code. Where custom code is needed (an API without an off-the-shelf connector, complex data transformation), we write and host it ourselves. Your engineers are looped in for credentials and for the security review, and that's it.
How long does a typical first build take?
From signed contract to live workflow: usually two weeks. The first half is the audit and design (we don't build anything until you've signed off the process map and the success metric). The second half is build, test, document, and hand over.

Find out what your team's manual hours are actually costing.

A two-day paid audit gives you a process map, a prioritised list of automations ranked by ROI, and a fixed quote for the build. £1,500 / $2,000. Roll the audit fee into the build if you decide to proceed.