Aircon and Facilities Servicing: Automating the Recurring Contract Schedule Before Q4

If you run an aircon servicing or facilities maintenance company in Singapore, the highest-return automation available to you is not a new CRM and not a customer-facing app. It is making your recurring contract book generate its own jobs. A 200-contract book on quarterly servicing is 800 visits a year, and every one of those visits needs a due date, a confirmed slot, an assigned technician, a completed service report and an invoice. Most SMEs in this trade still run that loop on a spreadsheet plus one coordinator's memory. That is why missed visits, double-booked crews and unbilled completed jobs never show up as an obvious failure — they show up as thin margin and a contract renewal that quietly does not happen.

Why does a servicing business lose money on the schedule rather than on the job?

The job itself is the part you have already priced correctly. You know how long a fan coil cleaning takes, what chemical wash costs you, and what a two-man crew bills per day. The leakage sits in the gaps between jobs:

• A contract visit that lapses past its due quarter — revenue you already sold and collected against, delivered late or not at all.
• A visit rescheduled by the customer that never gets rebooked because it lives in a WhatsApp thread, not in a system.
• A completed job whose service report sits in a technician's phone for three weeks, so the invoice goes out a month late.
• A crew sent to Jurong in the morning and Bedok at 2pm because nobody clustered the route.

None of these appear in your P&L under a line called scheduling. They appear as lower revenue per contract and higher labour hours per visit. Put rough numbers on it: 200 contracts at four visits a year at $90 per visit is $72,000 of contracted servicing revenue. If 6% of visits slip out of the year entirely, that is roughly $4,300 gone — before counting the renewal conversations that get harder when a customer can name the quarter you skipped.

What does the manual scheduling loop actually cost each month?

Time the loop honestly for one month. In most servicing SMEs we look at, the coordinator's week contains: pulling the next-due list from a spreadsheet, calling or messaging each customer for a slot, chasing the ones who do not reply, assigning technicians against a whiteboard, printing or photographing job sheets, chasing service reports back from the field, keying completed jobs into the accounting system as invoices, then chasing payment.

For a two-to-four-crew operation that is typically 25 to 40 hours a month of pure coordination — most of a full-time salary spent on data movement rather than on customers or quotations. The important part is not the cost. It is that this work does not scale: adding a fifth crew adds coordination load faster than it adds billable capacity, which is why servicing businesses often stall at a size that has nothing to do with demand.

Which part of the workflow should you automate first?

Order matters more than tooling here, and the common mistake is starting at the wrong end — buying a technician app first, when the technicians are not the bottleneck. The sequence that works:

1. Turn the contract register into data. Customer, site, asset (each unit, with location and serial), contract start and end, visit frequency, contracted price per visit. This is the foundation and it is usually one week of cleanup.
2. Auto-generate the next-due job list. The system, not a person, decides what is due in the next 21 days and what is already overdue.
3. Automate the confirmation, not the conversation. An outbound WhatsApp message proposing a slot, with a reply that books it. Exceptions go to a human; the routine 70% does not.
4. Capture completion in the field. Photos, checklist, chemicals used, customer signature — captured once, at the site.
5. Generate the invoice from the completed job. No re-keying. The job record already contains everything the invoice needs.

Each step is useful on its own, which matters — you get value in week two, not in month six.

How do you build this without replacing everything you already own?

You do not need to move your accounting system. What is missing in most of these businesses is the middle layer: contracts, assets, visits and jobs. Your accounting package handles customers and invoices but has no concept of a fan coil unit on level 3 of a building that is due for a chemical wash in November.

Build that middle layer, then connect it outward. Asset-level tracking is what separates a real servicing system from a calendar: when a customer calls about a unit, you want the last three visits for that specific unit, not for the site. It is also what makes upsell conversations concrete — a compressor with a documented service history is a quotation, not a guess.

What about technicians who will not use software?

Assume they will not install an app and will not remember a password. Design for that: a link sent by WhatsApp that opens their assigned jobs for the day, three taps to close a job, camera upload for photos, tolerant of a bad signal in a plant room. If closing a job takes longer than filling in the paper sheet, adoption fails and you are back to chasing reports. Supervisors handle the exceptions; the system should never require the field to handle anything unusual.

What should be running before the Q4 peak?

The next 30 days are the window. Deepavali and the year-end period pull servicing demand forward, and the pre-Chinese New Year rush in January is the hardest week of your year to coordinate manually. A realistic sequence: contract register cleaned and loaded in weeks one and two, auto-generated due lists and WhatsApp confirmations live by week four, field completion and invoice generation by end of October.

Two Q4 items connect to this directly. If you are GST-registered, the job-to-invoice chain should end in a structured invoice, which is the practical entry point to InvoiceNow and PEPPOL readiness — worth designing in now rather than retrofitting. And the grant windows for PSG and SkillsFuture Enterprise Credit have year-end constraints worth checking against your own eligibility before you commit budget. Treat both as workflow design decisions and confirm the specifics with your accountant or the relevant agency.

Frequently asked questions

How many contracts do I need before this is worth automating?

Roughly 80 or more recurring contracts, or two or more crews. Below that, a well-structured spreadsheet with a disciplined coordinator genuinely works. Above it, the coordination load grows faster than the revenue and you start losing visits you have already been paid for.

Can I do this with WhatsApp and Google Sheets alone?

Partially, and it is a reasonable first month. What breaks is asset-level history, automatic due-date generation across hundreds of contracts, and the job-to-invoice handoff. The failure is usually silent — the sheet stops matching reality and nobody notices until a customer calls about a missed quarter.

Will a government grant cover this?

Pre-approved field service management packages appear on the PSG solution list, and custom-built systems typically fall under different schemes with different criteria. Eligibility depends on your company profile and the specific solution, so check the current listings on the official portal before assuming coverage.

If you want a view of where your own servicing schedule is leaking, start by pulling one number: how many contracted visits due in Q3 have not yet been completed. That figure usually settles the business case on its own.