AI Opportunity Report
Prepared for
Dave Okonkwo
Owner
Session
14 July 2026
62 minutes
Business
11 people
Wagga Wagga, NSW
Recommendations
9 actions
4 things not to do
Ridgeline is spending about 87 hours a month on work that does not need a person doing it. Thirty of those hours can be recovered this month for nothing, using software you already pay for.
Your office manager is the constraint, not your technicians. Six of the nine items below sit on her desk. Three of them are features already licensed inside ServiceTitan, Jobber and Xero that were never switched on — that is not a criticism, it is the most common finding in this industry, and it is the cheapest money in this report.
The single largest item is rebuilding the dispatch board by hand. That one needs building and it is the last thing you should do, not the first.
Start here · costs nothing · about one day of setup
Switch on estimate follow-up in Jobber, invoice reminders in Xero, and job-status notifications in ServiceTitan. Together those three recover 30 hours a month, worth $13,600 a year, at no additional software cost. Detail in items 01, 06 and 07.
This is the business as you described it in the session. If any of it is wrong, the numbers that follow are wrong too — read this part first and tell me.
| The team | You, one office manager (Steph, full time), six service technicians, two installers, one apprentice. |
| Volume | Around 210 inbound calls a month. About 40 estimates out the door, average $2,900. Six to eight installs a month, average $9,500. |
| Software you pay for | ServiceTitan (jobs, dispatch, invoicing), Jobber (estimates, retained from before the migration), Xero, Google Workspace, a shared mobile number, one WhatsApp group per crew. |
| Where the day actually goes | Steph starts at 6:45am rebuilding the board for the day. Customers ring through the morning asking where their technician is. You personally chase estimates on Friday afternoons when you remember. |
| What only you can do | Quoting anything over $8,000, hiring, supplier relationships, and any job where the customer is unhappy. |
| What you said you'd fix first | "Stop Steph drowning before 8am." Noted — item 04 addresses it, but it is not where I would start. |
Ranked by value against effort — not by size. Hours and dollars in this table are labour only: time a person currently spends that they would not have to. Where there is also a revenue effect, it is called out in the detail below and deliberately excluded from these totals, because I cannot prove it and I am not going to pad the number with something I cannot prove.
| # | Action | Buy or build | Cost | Hrs/mo | $/yr |
|---|---|---|---|---|---|
| 01 | Estimate follow-up | Jobber — already paid | $0 | 9 | $4,100 |
| 02 | Missed-call text-back | Buy — OpenPhone + AI | $89/mo | 6 | $2,700 |
| 03 | After-hours enquiry triage | Same tool as 02 | $0 extra | 7 | $3,200 |
| 04 | Morning schedule prep | Build — 3 weeks | $6,800 | 24 | $10,900 |
| 05 | Install readiness & permits | Build — 2 weeks | $4,200 | 11 | $5,000 |
| 06 | Invoice reminders | Xero — already paid | $0 | 8 | $3,600 |
| 07 | "Where's my technician?" | ServiceTitan — licensed | $0 | 13 | $5,900 |
| 08 | Maintenance-plan renewals | Build — small | $1,900 | 5 | $2,300 |
| 09 | Job photo filing | Buy — CompanyCam | $29/mo | 4 | $1,800 |
| Total labour recoverable | 87 | $39,500 | |||
Now: Forty estimates leave each month. Follow-up happens when you remember on a Friday, which by your own estimate is two Fridays in four. You rebuild the list by hand each time from Jobber's open-estimates view.
Do this: Jobber's quote follow-up automation is included in your plan and is currently off. Turn on a three-touch sequence — day 2, day 7, day 14 — from Steph's address, in your voice. Any reply goes to her, not to a bot. Nothing is sent to a customer you have flagged.
Cost and time: Nothing. About an hour to configure, most of it spent writing the three messages properly.
Revenue effect, excluded from the totals: Your current estimate close rate is around 22%. Consistent follow-up moves that in most shops. I am not going to put a number on your business without measuring it, so here is the test instead: run it for eight weeks and compare close rate on estimates issued in that window against the previous eight. If it does not move, you have lost an hour of setup.
Now: Around 210 calls a month come in. Steph estimates she cannot reach roughly 15% of them — call it 31 calls. Most leave no voicemail. She works back through the missed-call log in gaps between other work.
Do this: Move the shared mobile number onto OpenPhone. Any call that rings out gets an SMS within ten seconds: who you are, and one question about what they need. The reply lands in a shared inbox with the caller's number and history attached. Steph stops reconstructing the log; she works a queue.
Cost and time: $89/mo for the plan and AI add-on. Half a day to port the number and write the message. Keep your existing ServiceTitan setup untouched.
Watch for: Do not let it answer questions about price or availability. It asks what they need and confirms a human will call. Anything more and it will make a commitment you cannot keep.
Now: Evening and weekend calls come to your personal phone. You said you answer "most of them", and that Sunday is the worst. There is no record of the ones you handled verbally, so Steph re-asks the customer on Monday.
Do this: The same OpenPhone setup, on an after-hours rule. Genuine emergencies — no heat, no cooling, water — ring straight through to whoever is on call. Everything else gets the text, is logged, and is first in Steph's queue at 7am with the customer's own words attached.
Cost and time: Included in item 02. An hour to define what counts as an emergency, which is a decision only you can make.
Now: Steph starts at 6:45am. She reconciles overnight calls, technician texts from the WhatsApp groups, yesterday's incomplete jobs and today's booked work into a board that holds until about 9:30am, when the first change lands. Roughly 70 minutes a day, plus 30 to 40 minutes of re-work through the morning.
Do this: A coordination layer that reads ServiceTitan, the WhatsApp crew groups and the OpenPhone inbox from item 02, and produces a single draft board with every conflict flagged and a plain-English note on each. Steph approves or overrides. It does not dispatch anybody. She keeps every decision; what she stops doing is assembling the picture.
Cost and time: $6,800, three weeks. Payback around seven months on labour alone.
Why this is last, not first: It is the biggest number in the report and it is the item you most want fixed. It also depends on items 02 and 07 being in place — without them it would be reading an incomplete picture, and a draft board built on incomplete information is worse than no board. Do the free items first, then this.
Now: Six to eight installs a month. Equipment, permit, inspection booking, customer availability, finance approval and the job pack all have to land in order. You told me two crews in the last quarter arrived at jobs that were not ready. Nobody can see the full reason a job is held; they can only see that it is.
Do this: A readiness record per install: every check, who owns it, the date it is needed by, and the evidence. It pulls what ServiceTitan already holds, chases the gaps by message, and will not let a job move to scheduled until every check is green or you override it with a reason.
Cost and time: $4,200, two weeks. The labour saving understates this one — a crew standing around on a $9,500 install costs more than Steph's afternoon.
Now: Steph runs an aged receivables report on Wednesdays and emails overdue accounts by hand. Anything past 45 days comes to you.
Do this: Xero's invoice reminders, which you pay for and have never enabled. Three touches at 7, 14 and 30 days overdue. Exclude your four commercial accounts on payment terms — those need a phone call from you, not an email.
Cost and time: Nothing. Twenty minutes in settings, plus writing the three emails.
Now: The single biggest interruption in the office. Steph fields these all morning, and each one requires her to check the board, and sometimes to text the technician.
Do this: ServiceTitan's customer notifications are part of your licence and are switched off. Turn on the on-the-way and delay notifications with an arrival window. Customers stop ringing because they already know.
Cost and time: Nothing. Half a day, most of it spent getting the arrival windows honest rather than optimistic. A window you miss is worse than no window.
Now: You have 340 maintenance-plan customers. Renewal reminders go out when Steph notices the anniversary in a spreadsheet she maintains separately from ServiceTitan.
Do this: Retire the spreadsheet. Drive renewals off the ServiceTitan record with a reminder at 45 and 14 days, and a booking link for the service visit in the same message.
Cost and time: $1,900, about a week. Do this after item 04 — same underlying plumbing.
Now: Technicians photograph work on their own phones and send it to the crew WhatsApp group. Steph pulls the images out and files them against the job when she needs them for a warranty claim, which is when she discovers the ones nobody sent.
Do this: CompanyCam, which files against the job automatically from the technician's phone and tags location and time. Lowest-value item here, but it is nine dollars a head and it removes a warranty risk.
These came up in the session or are the obvious next thought. I do not think any of them are worth your money right now, and I would rather say so here than have somebody sell you one.
Do not replace ServiceTitan.
You said it "doesn't do what we need". Three of the nine items above are features you are already paying for inside it. Switch them on, live with it for a quarter, and then decide. A migration would cost you more in disruption than everything in this report saves.
Do not put an AI chat widget on the website.
Four enquiries a month arrive through the site. Two hundred and ten arrive by phone. Fix the phone.
Do not automate technician job notes yet.
Those notes are your evidence on a warranty dispute. The problem today is that they are inconsistent, and automating an inconsistent input gives you consistent nonsense. Agree what a note must contain, run it on paper for a month, then automate it.
Do not let anything quote a customer.
You told me every job over $8,000 goes through you. Keep it that way. Nothing in this report prices work, commits to a date, or answers a complaint.
| When | Items | You spend | Hrs/mo back | $/yr back |
|---|---|---|---|---|
| This month One day of Steph's time | 01 estimate follow-up 06 invoice reminders 07 technician notifications | $0 | 30 | $13,600 |
| Next 60 days Two tool subscriptions | 02 missed-call text-back 03 after-hours triage 09 job photo filing | $118/mo | 17 | $7,700 |
| When you're ready Six weeks of build | 04 morning schedule prep 05 install readiness 08 renewal reminders | $12,900 | 40 | $18,200 |
Do not do all three stages at once. The first stage costs nothing and returns a third of the total — if that lands and Steph's morning is measurably better, you will know whether the rest is worth it. If it does not land, something in section 2 was wrong and we should talk before you spend anything.
You do not have to. Items 01, 06 and 07 are settings — Steph can do them this week from the detail above. Items 02, 03 and 09 are sign-ups. Only three items need building, and any competent developer can work from this document.
If you want me to do it, this is the price, fixed, and it does not change after we start:
| 04 · Morning schedule prep — 3 weeks | $6,800 |
| 05 · Install readiness & permits — 2 weeks | $4,200 |
| 08 · Renewal reminders — 1 week | $1,900 |
| All three, six weeks | $12,900 |
Half up front, half on delivery. Thirty days of fixes included after each item goes live. If a build does not do what this document says it will do, you do not pay for that item.
Every hour figure above is an estimate built from what you and Steph told me in a 62-minute session. None of it is measured. You should treat the ranking as reliable and the individual numbers as approximate, and you should argue with any of them.
| Office labour rate | $38/hr, fully loaded. Applied to all nine items, since all nine sit with Steph or you. |
| What "recoverable" means | Time currently spent that would not need to be spent. It does not mean you will reduce headcount — it means Steph gets her mornings back, which is what you said you wanted. |
| Excluded on purpose | Revenue from better follow-up, revenue from rescued missed calls, and the cost of a crew standing idle. All three are real and all three are unprovable from a single conversation, so none are in the $39,500. |
| Tool prices | Public list prices as at 17 July 2026, for your seat count. I take no commission from any vendor named in this document. |
| What would change the answer | If your real missed-call rate is closer to 5% than 15%, item 02 drops several places. If you migrate off Jobber, item 01 needs rebuilding elsewhere. If Steph leaves, everything in section 6 becomes urgent rather than optional. |
This report is yours. Use it with your team, give it to another provider, or ignore it. There is no exclusivity clause and no obligation to buy anything else from me.
Questions on any number in here — call me. Liam Deetlefs · support@comora.co · comora.co