iRepair Media — homeiRepairMedia

Connect

Automation & Integrations

Most businesses do not need more software. They need the software they already pay for to talk to itself. Automation is the least glamorous work we do and frequently the highest return, because it removes hours that recur every week.

What this fixes

The same data is entered repeatedly

An enquiry typed into the CRM, then the quoting tool, then the accounting package. Three chances to introduce an error and hours lost every month.

Leads sit unattended

A form submission arriving in a shared inbox that nobody owns. Response time is often the single biggest factor in whether an enquiry converts.

Nobody knows the real numbers

Enquiries in one system, jobs in another, invoices in a third. Answering a simple question takes an afternoon of exports.

Manual steps that only one person understands

A weekly routine that exists entirely in someone's head. That is a risk, not a process.

What’s included

Process mapping
How work actually flows today, including the undocumented steps. Automating a broken process just makes it fail faster.
System integration
APIs and webhooks connecting website, CRM, email, accounting and scheduling so records stay in step.
Lead routing
Enquiries captured, categorised and sent to the right person immediately, with a fallback if nobody responds.
Automated communication
Acknowledgements, reminders and follow-ups that go out reliably and still read as though a person wrote them.
Reporting
One place showing enquiries, sources and outcomes, without anyone exporting a spreadsheet.

Integrations

Connecting existing tools is usually cheaper and safer than replacing them. Where a product offers a decent API we use it; where it does not, there is normally a supported route via webhooks or an integration platform. We prefer connections you can understand and we document what talks to what, so a change in one system does not silently break another.

  • Website forms into your CRM, with source and campaign preserved
  • CRM to accounting for quotes, invoices and payment status
  • Calendar and scheduling for bookings and job allocation
  • Stock and fulfilment for ecommerce operations
  • Notifications into the channels your team already watches

AI workflows

We use AI where it does a specific job measurably well, and we are straightforward about where it does not. It is genuinely good at summarising, drafting, classifying and extracting structure from messy text. It is unreliable as an unsupervised decision-maker in anything that touches money, safety or a customer relationship. Every workflow we build keeps a person in the loop at the point where being wrong would actually cost you something.

  • Summarising and categorising inbound enquiries
  • Drafting first-pass replies for a human to approve and send
  • Extracting structured data from documents and forms
  • Internal search across your own documentation
  • Clear boundaries on what runs without a person checking it

Ecommerce automation

A shop generates events all day: a basket left open, a first order, a parcel leaving, a customer who has not been back since spring. Each is a reason to send something specific, and most of it can go out without a person deciding to. We build the sequences on the store's own order and customer records in WooCommerce, with FluentCRM handling the messaging, so what a customer receives matches what actually happened to their order. A checkout that loses people structurally is a build problem rather than a messaging one, and the WooCommerce page deals with that side of it.

  • Abandoned basket sequences triggered from live cart data, spaced rather than instant
  • Welcome sequences that introduce the range instead of leading with a discount
  • Post-purchase journeys covering dispatch, use, review requests and replenishment
  • Win-back campaigns aimed at customers who ordered once and did not come back
  • Order, payment and dispatch notifications driven by order status in WooCommerce
  • Segmentation by what someone bought, how recently and how often

Follow-up after the enquiry

Capture and routing are covered further up this page. This section is about the fortnight afterwards. An enquiry arrives carrying facts worth keeping — which service it was about, which town, whether it came from search, an ad or a referral — and tagging those as it lands is what lets follow-up differ later, because a survey request and a parts enquiry should not receive the same three messages. The sequence is written once and then runs on its own, and a reply from the enquirer stops it. Enquiries that go quiet move to a slower list and are contacted again when there is something worth saying. On WordPress this runs in FluentCRM; where the CRM sits elsewhere, the same sequences run over its API.

  • Tags for service, location and source applied on arrival, so later messages can differ
  • Follow-up that continues past the first attempt and stops when someone replies
  • Quote chasing on a schedule rather than when somebody remembers
  • Re-engagement for older enquiries, sent when there is a genuine reason to

Bookings and rentals

The form is the part a customer sees. What decides whether a booking can be trusted sits behind it: what is genuinely free, how far ahead someone may commit, how long a job blocks out, and what it costs once the options are chosen. Those rules go into the system rather than staying with whoever answers the phone, so the price is worked out at the point of booking and availability is checked before a slot is offered rather than after it is taken. Blyth Storage, Meadow View Storage and Newburn Container Storage each run automated online storage rental, where a customer chooses a unit and starts a rental without anyone taking the booking by hand.

  • Booking forms that collect what is needed to schedule and price the job
  • Availability rules: lead times, blackout dates, capacity and duration
  • Pricing logic applied at the point of booking rather than quoted afterwards
  • Payment taken at the point of commitment where the booking warrants it
  • Confirmations and reminders sent from the booking record rather than by hand
  • Automated storage rental at Blyth Storage, Meadow View Storage and Newburn Container Storage

Stock, listings and pricing kept in step

Information that has to appear in more than one place is usually kept in step by somebody typing it twice: stock levels, vehicle or property listings, availability, trade prices, a catalogue that changes weekly. Two copies of the same list will drift, and reconciling them becomes a standing task nobody was hired for. The work is to agree which system owns each set of records and have the website read from that one, so a listing is created once, where the business already manages it. Bebside Van Centre works this way — the vans are listed on eBay, and the website pulls those listings through the eBay API rather than the same stock being maintained in two places.

  • One agreed owner per set of records — stock, listings, availability, pricing
  • Live eBay vehicle listings pulled onto the Bebside Van Centre site via the eBay API
  • Refresh cadence set per source, scheduled with cron where a fixed interval suits it
  • Mapping rules for fields that do not match: units, categories, tax-inclusive against tax-exclusive prices
  • The copy that is not the source treated as display only, so it is not edited into disagreement
  • Failure handling agreed up front: stale or malformed data flagged rather than shown as current

Operational workflows

Part of the work sits behind the website entirely: a status that has to be changed in two systems, a document produced by hand, a Monday routine that lives in one person's memory. The mechanical parts of that can move into the system — a status change that triggers the next step, a task that runs to a schedule with cron, an internal screen built around the job somebody does all day rather than around whichever plugin happened to provide it. Blyth Storage has a bespoke Vercel dashboard for its search and project work. Where an internal tool grows past a few screens and scheduled tasks into an application with its own data model, that is Laravel work.

  • Status changes that trigger the next step instead of waiting to be noticed
  • Recurring tasks run to a schedule with cron rather than remembered
  • Internal screens built for the handful of jobs a team repeats every day
  • Documents and notifications generated from the record rather than retyped
  • A bespoke Vercel dashboard for search and project work at Blyth Storage

Campaign automation

The sequences above are keyed to an order or an enquiry. This is the rest of it: the campaigns a business sends because it has something to say, and the decision about who should hear it. Segmenting is the part worth spending time on — by service interest, by stage, by how recently somebody engaged — because that is what lets a message be narrower than the whole list. Funnels can run continuously rather than in bursts: a contact qualifies, their sequence starts, and it carries on whether or not anyone is at a desk that week. Consent and preferences live in the same records as everything else rather than in a separate list that goes out of date.

  • Segments built from service interest, stage and recency, held in the CRM
  • Evergreen funnels that start when someone qualifies rather than only on send day
  • Suppression rules configured so the same person is not sent two things in a day
  • Campaign scheduling handled in FluentCRM alongside the customer record
  • Consent and preference handling built into the record rather than added afterwards
  • A clear split between transactional messages and marketing ones

Questions

Where should we start?
With whichever manual task is repeated most often. It is usually moving enquiry details between two systems, or chasing something that could remind itself. Start narrow, prove the time saved, then extend — large automation projects that try to fix everything at once tend to stall.
Do we need to replace our current systems?
Usually not. Most of this work is connecting what you already run. Replacing a system your team knows is a much bigger undertaking than integrating it, and we would only suggest it if the tool genuinely cannot do the job.