HomeServicesAIAutomations

Less admin, more real work.

We automate the repetitive tasks that fill your week, from sorting enquiries to updating records. AI helps with the steps that need reading or judgement, and a person approves anything customers see.

  • Less copying and pasting
  • Enquiries routed quickly
  • Every step logged

Supplier invoices

TriggerInvoice arrives by email
TaskAI reads and extracts details
ReviewFinance approves the draft
TaskBill created in Xero

Small teams lose hours each week copying details between inboxes, spreadsheets and systems that never quite connect. We map those tasks, automate the safe ones, and add a human check wherever a mistake would matter.

Tasks worth automating first.

We start with tasks that repeat often and follow clear rules, then add AI where a step needs reading or judgement.

Enquiry triage

New emails and form entries are read, tagged by type and urgency, and sent to the right person with a short summary attached.

Drafted replies

AI drafts replies in your tone using your FAQs and past answers, and they wait in a queue for a person to approve.

Document data capture

Details from invoices, receipts, order forms and CVs are pulled into your accounts or CRM, with uncertain fields flagged for checking.

CRM updates

Contacts, deals and notes in HubSpot or your CRM are created and updated from emails, calls and forms without retyping.

Reports and summaries

Weekly summaries of sales, enquiries or support tickets are compiled from your tools and sent by email or Slack.

Errors and logs

Every run is logged, and failed steps send an alert with enough detail for someone to fix and rerun them.

Done properly.

The details that decide whether automations pays off.

The right tool for each workflow.

Zapier is quick to set up and easy for your team to edit. Make handles more complex branching, and n8n can be self-hosted when data needs to stay on your own servers. We pick based on volume, cost and control.

Automation toolkit

ZapierMaken8nOpenAIClaudeGmailMicrosoft 365HubSpotXeroSlackGoogle SheetsAirtable

Unsure items go to a person.

When the AI is not confident about an item, it stops and waits for review instead of moving on. You choose which steps run on their own and which wait for approval, and you can change that as trust builds.

Workflow settings

Approval step

Before customer emails

Low confidence

Sent for review

Failed runs

Alert in Slack

Run history

Kept and searchable

How it works.

Clear steps, and you approve each one before we move on.

  1. 01

    Audit

    We list repetitive tasks, how often they happen and the tools involved.

  2. 02

    Prioritise

    We rank tasks by time saved, risk and ease of building.

  3. 03

    Build

    We build each workflow, add review steps and test with real examples.

  4. 04

    Monitor

    We watch the logs, fix failures and tune prompts as your work changes.

Example project

A recruitment agency buried in CVs and enquiries

A small recruitment agency receives CVs and client enquiries across several inboxes. Consultants copy details into their CRM by hand, replies are slow, and good candidates sometimes get missed.

What we would do

  • Map how CVs and enquiries arrive today
  • Extract candidate details into the CRM automatically
  • Draft replies for consultants to approve
  • Send a morning summary to Slack

An illustration of how we would approach it, not a client result.

What you would get

  • Tested workflows in Make or n8n
  • Candidate records created from CVs
  • Drafted replies waiting for approval
  • Run logs and failure alerts
Start a project like this

Questions, answered.

What is the difference between automation and AI automation?

Standard automation follows fixed rules, such as copying a form entry into a spreadsheet. AI automation adds steps that need reading or judgement, such as sorting an email by intent or drafting a reply for someone to approve.

Is it safe to send our data through these tools?

We agree which data each workflow can touch and keep sensitive details out where possible. We use business accounts where your data is not used to train models, n8n can be self-hosted, and every setup follows UK GDPR.

Can my team change the automations later?

Yes. We build in tools your team can open and edit, document each workflow, and show people how to make simple changes. Larger changes can be handled by us on a fixed monthly fee.

How much does an automation project cost?

It is quoted after a short call, with a fixed price agreed before we start. Tool subscriptions such as Zapier or Make are paid directly, and ongoing improvements can run on a fixed monthly fee.

Hand the busywork to automation.

Tell us what you need. Charlie will come back with a plan and a price, usually within one working day.

hello@infinity.agency