The Full Audit — AltaPoint Development
Your whole operation, measured, in writing.
$999 one time
Not an estimate from a short form. A deep report built from your actual records — every leak named, priced in dollars and hours, put in fix order, with a walkthrough call included to go through it together.
The report is yours to keep and use with anyone.
The guarantee: If we can’t document 5 hours a week in savings, the audit is free.
What you get
- The written Full Audit report, built in the fixed anatomy below
- One walkthrough call to go through the report with you, finding by finding
- A ranked fix order you can work yourself, hand to another builder, or bring back to us
This is a standalone product. It is not a step you have to take before anything else, and buying it does not obligate you to buy anything after it.
The anatomy — the same order, every time
- Verdict first. The headline findings up top, not buried: total findable leaks in dollars per year, total time in hours per week, and the single biggest leak named.
- The leak registry. Every leak we find, named in a table, with what we measured, what it costs in dollars per year, and what it costs in hours per week.
- Counts and drivers. How many leaks, of which kind — calls, follow-up, money, scheduling, admin — and what is driving each kind, in plain words.
- Ranked fix order. What to fix first, second, third — and what each fix pays back.
- Do-it-yourself vs. hire-it-out. An honest split: what your shop can handle in-house, including settings in software you already pay for, and what needs a builder.
- What we could not determine. Stated plainly. Missing data counts against the picture — it is never smoothed over to make the report look cleaner. No dollar figure is attached to anything in this section.
- The recommended tool stack. The existing tools behind the fixes — what each one fixes, its published price, and what it takes to set up — with the stack’s total cost set against the leaks it addresses. You buy them direct; we don’t sell tools or take a cut.
- Sources appendix. Where every number came from: which record, which dates, and how the figure was worked out. Any rate used in the math is listed here and labeled as an assumption where it is one. Tool prices trace to the vendors’ published pricing pages, each listed with its check date.
How it works
- You buy the Audit — $999 one time.
- We send the data list the same business day. Call logs, estimate records, invoicing records, lead and follow-up records if you keep them, and your current software list.
- You send what you have. What you do not have is recorded as missing — missing is a finding, not a delay. Missing items are requested at most twice.
- We build the report from your records in the anatomy above. Every figure traces to a source in the appendix.
- You get the written report, then the walkthrough call. We go through the verdict, the leaks, and the fix order together, and you ask anything you want about any figure.
This is the actual document the $999 buys.
A Full Audit report — verdict first, every leak named in dollars and hours, the fix order, the recommended tool stack with prices, and what we could not determine, stated plainly. Below is a complete sample for a fictional shop — every section of the real document, end to end — so you can read exactly what the $999 buys before you pay anything.
FULL AUDIT — Harborline Plumbing
1Verdict first
Harborline has about $78,500 a year in findable leaks, costing 10.5 hours a week of owner and office time.
The single biggest leak is after-hours calls: $41,225 a year — more than half the total. Nearly two out of every three missed calls come in after hours or on weekends, when nobody is there to answer, and every one of those callers had a plumbing problem and a phone in their hand.
None of the five leaks below needs new software to start fixing. Two of the five fixes are true AI — an agent that talks to callers, and one that reads your estimates and drafts the follow-up. The other three are plain automations and settings in Jobber and QuickBooks Online, tools Harborline already pays for — and the fix order labels them that way, because they are not AI and do not need to be.
2Leak registry
| # | Leak | What we measured | $ / year | Hours / week |
|---|---|---|---|---|
| L1 | Missed after-hours calls | 244 missed after-hours/weekend calls a year × 35% would book = 85 jobs × $485 average job value | $41,225 | 2.0 (owner) |
| L2 | Estimates that go quiet | 176 estimates a year get no follow-up. Followed-up estimates close at 41%; unfollowed close at 18% — a 23-point gap ≈ 40 jobs × $485 | $19,400 | 1.5 (owner) |
| L3 | Invoices paid late or never | Write-offs of $1,455 in the window × 4 quarters. 54 of 360 invoices (15%) were paid more than 14 days late; average days-to-pay is 17 | $5,820 | 1.5 (owner) |
| L4 | Scheduling & reminders by hand | Office time spent booking, confirming, and reminding by hand: 4.0 hr/wk × $30/hr × 52 weeks | $6,240 | 4.0 (office) |
| L5 | Duplicate entry, Jobber ↔ QuickBooks | The same job and invoice details typed twice, once in each system: owner 1.5 hr/wk × $75/hr × 52 weeks | $5,850 | 1.5 (owner) |
| TOTAL | Total time split: owner 4.5 hr/wk + office 6.0 hr/wk = 10.5 hr/wk | $78,535 / year | 10.5 hr/wk |
How to read the hours column: for L1, L2, and L3 the hours are the owner's time spent living with the leak today — covering the phone, chasing estimates, and chasing payments. For L4 and L5 the hours are the time the manual work itself takes each week.
3Counts and drivers
5 leaks, one of each kind:
| Kind | What it means, plainly | Leak |
|---|---|---|
| Calls | Calls that come in and never get answered | L1 |
| Follow-up | Estimates sent, then never chased | L2 |
| Money | Invoices paid late, or never paid at all | L3 |
| Scheduling | Booking, confirming, and reminding done by hand | L4 |
| Admin | The same information typed into two systems | L5 |
Two of the five leaks — calls (L1) and follow-up (L2) — are the same problem wearing two hats: a customer reaches out, and nobody gets back to them in time. Together they are $60,625 of the $78,535 — about three-quarters of everything we found. That is why the fix order below starts with getting back to people.
Drivers, in plain words:
- Coverage. The owner is the phone. When he is on a job, driving, or off the clock, calls ring out — 96 missed calls in 90 days, 61 of them after hours or on weekends.
- The owner is the follow-up department. Deciding which estimates are worth chasing takes the owner's judgment, and he is on jobs all day — so 44 of 118 estimates got no follow-up at all. Followed-up estimates closed at 41%; the rest closed at 18%. The gap is not price or workmanship — it is a judgment bottleneck.
- Payment terms live in people's heads. Invoices go out without a reminder path or a card on file, so payment waits until someone chases it — 15% land more than 14 days late, and 3 were written off in the window.
- Hand work that software already does. Confirmations, reminders, and the Jobber-to-QuickBooks handoff are all done by hand, even though both systems already have the settings to do them.
4Ranked fix order
Fix 1 — An after-hours agent that actually talks (fixes L1 · $41,225/yr)
What kind: true AI. Root cause: no coverage — nobody is there, and what the caller needs is a conversation, not a recording. An AI agent answers after-hours and weekend calls in Harborline's name: it works out whether it is an emergency or can wait, gives the shop's standard prices for the common jobs, books real jobs straight onto the calendar, and texts the owner a summary of every call. A missed-call text-back stays on as the backstop for any call the agent cannot take — that part is a plain automation, and it is labeled that way.
Steps: load the agent with Harborline's price list, booking rules, and service area → write its opening and escalation rules in the owner's words (anything safety-related, or an angry customer, goes straight to the owner's phone) → set the coverage hours (after 5 PM and weekends) → run test calls, including a burst-pipe scenario, before it goes live.
Payback: the largest leak by more than two to one. Recovering even a small share of 85 lost jobs a year covers the cost of coverage many times over. One recovered job is $485.
Fix 2 — Estimate follow-up that reads first (fixes L2 · $19,400/yr)
What kind: true AI, with one automation inside it. Root cause: judgment bottleneck — the owner is the only one who can tell which estimates deserve a chase, and he is on a job all day. So the fix is not a louder reminder: an agent reads every open estimate with its job notes and contact history, ranks the ones worth chasing this week and says why, and drafts each follow-up in the owner's voice — a different message for the repipe than for the water heater. The owner approves each one with a tap before it sends. Underneath it, the plain day 2 / day 5 / day 12 nudge still runs in Jobber for small standard estimates that just need a touch — that part is automation, and it is enough for those.
Steps: connect the agent to Jobber's estimates and job notes → set the ranking rules with the owner (age, size, last contact, job type) → load the plain scheduled nudge for small standard estimates → the owner works a short approval queue, about ten minutes a day → review the rankings together weekly for the first month.
Payback: closing the 23-point gap between followed and unfollowed estimates is worth about 40 jobs a year. The agent does the reading and the drafting; the owner's judgment stays on every message that goes out.
Fix 3 — Automatic confirmations & reminders (fixes L4 · $6,240/yr)
What kind: automation — not AI, and it does not need to be. Root cause: pure volume — the same confirmations and reminders, the same way, hundreds of times a year. Turn on automatic booking confirmations and appointment reminders, plus a review ask after each finished job.
Steps: turn on confirmations and reminders in Jobber → add the review ask to the job close-out → hand-check the first week of messages before trusting them.
Payback: gives back 4.0 hours a week of office time — the biggest block of hours in this report — for settings work only. No new software.
Fix 4 — Payment reminders + card-on-file (fixes L3 · $5,820/yr)
What kind: automation, plus one process fix. Root cause: no consistency — payment terms lived in people's heads, so every invoice got chased a different way, or not at all. Automatic payment reminders on a set schedule, and a card-on-file line on the estimate so payment terms are agreed before the work starts.
Steps: set the reminder schedule in QuickBooks Online (before due, on the due date, 7 days after) → add the card-on-file line to the estimate template → make sure online payment is switched on.
Payback: aimed straight at the write-offs — $1,455 last quarter alone — and at pulling the average days-to-pay down from 17. Settings in QuickBooks Online Harborline already pays for.
Fix 5 — Kill duplicate entry: turn on Jobber ↔ QuickBooks sync (fixes L5 · $5,850/yr)
What kind: automation — a connection between two systems, nothing more. Root cause: pure volume — the same details typed twice, every job, every week. Turn on the sync between Jobber and QuickBooks Online so a job is typed once and appears in both.
Steps: connect the sync → match the fields between the two systems → hand-check the first week of synced records → stop typing anything twice.
Payback: gives back 1.5 hours a week of owner time at $75/hr and removes a steady source of mismatched numbers between the two systems. It is a setting in software already in the shop — it is last only because the dollar figure is the smallest of the five, not because it is hard.
5Do it yourself vs. get help
Do it yourself — fixes 3, 4, and 5. These three are settings in software Harborline already pays for. A careful afternoon in Jobber and QuickBooks Online, working from this report, gets all three done:
- Fix 3 — automatic confirmations & reminders (Jobber settings)
- Fix 4 — payment reminders + card-on-file (QuickBooks Online settings + one line on the estimate template)
- Fix 5 — Jobber ↔ QuickBooks sync (turn it on, then check the first week of synced records by hand)
Get help with — fixes 1 and 2. These two touch how customers are answered and what gets said to them, and they have to be set up so the owner's voice and judgment stay in them:
- Fix 1 — after-hours call coverage
- Fix 2 — estimate follow-up sequence
Help here can be us, or another builder. This report is complete enough to hand to either. The tools behind the do-it-yourself fixes are priced in §7 — every one of them is something you buy yourself, direct.
Where AI doesn't belong in this shop
Part of this report's job is saying where we would not put AI — even though AI is what we sell. For Harborline, that is three places:
- The final price on custom work. The agent can pull what the last three repipes cost and what they billed — and then the owner prices the job. A crawlspace he has seen beats a pattern in the records, and his name is on the number.
- Bad-news and dispute conversations. The "we opened the wall and found galvanized everywhere" call, and the customer who says the work was not done right — those are the owner's voice, live. The agent can draft his talking points and pull the job history first. It does not have the conversation.
- Safety calls. Whether a water heater with heat damage is safe to leave in service overnight is a licensed tech's call, on site, every time. No agent output goes anywhere near it.
6What we could not determine
These items are real. We are listing them instead of guessing at them, and no dollar figure is attached to any of them.
- Reviews that were never requested. Harborline finished 360 jobs in the window and received 11 new Google reviews. There is no record of how many customers were ever asked, so we cannot count the reviews that were missed — and we will not put an honest dollar figure on reviews that were never asked for.
- Referral leakage. Referrals are one of Harborline's lead sources, but referrals are not tracked anywhere — not in Jobber, not in the call log. Work that came in by word of mouth and work that was lost cannot be separated, so referrals cannot be counted at all.
- The 35 daytime missed calls. Of the 96 missed calls, 61 were after hours or weekends and 35 were daytime. The call log does not show whether a daytime caller was a new lead or an existing customer calling back. Only the 61 after-hours misses are counted in L1. That makes L1 a conservative figure — the real number is likely higher, and we are reporting the lower one.
7The recommended tool stack
Everything in fixes 2 through 5 runs on four tools. Two of them — Jobber and QuickBooks Online — are already in the shop; for those, the recommendation is the plan tier that carries the feature, not a new purchase. The other two are new. Buy these direct from the vendors — we don't sell them, we don't take a cut, and there is no AltaPoint markup on any of them. This section stays useful even if you never hire us: it is the shopping list behind the fix order, with real prices.
One leak is deliberately not in this stack: L1, after-hours calls. Fix 1 only works as an agent loaded with Harborline's own prices, booking rules, and words — the off-the-shelf answering tools we reviewed recite a generic script and know none of that. So Fix 1 is a setup job, not a purchase — it sits in §5 under "get help" for that reason.
| Tool | What it fixes | Plan | Price | Setup |
|---|---|---|---|---|
| Jobber Already in the shop | L2 + L4 — the estimates, notes, and schedule Fix 2's agent reads and drafts from, plus its plain scheduled nudges; online booking, confirmations, reminders (Fix 3) | Connect (team of 6–10) | $229/mo | 2–3 hours |
| QuickBooks Online Already in the shop | L3 + L5 — payment reminders and card-on-file (Fix 4); the other end of the sync that ends double entry (Fix 5) | Essentials | $85/mo | 1–2 hours |
| NiceJob | The review ask in Fix 3 — and the reviews gap in §6 that we would not put a dollar figure on | Starter | $95/mo | 1 hour |
| CompanyCam | Supports L3 — time-stamped job photos that settle "was it finished?" before it stalls an invoice; photos also ride along with the Fix 2 follow-ups | Crew (5 users) | $177/mo | 1 hour |
| The stack, at list prices | Addresses L2, L3, L4, L5 and the review ask — $37,310 of the $78,535 | $586 |
Jobber — Connect (team of 6–10) (already in the shop)
What it fixes: L2 and L4. Jobber is where the estimates, job notes, and schedule live — Fix 2's agent reads them from here and drafts from here, the plain day 2 / day 5 / day 12 nudge for small standard estimates is a Jobber automation, and Fix 3's online booking, confirmations, and reminders are Jobber settings. What it replaces: the office's mental list of who got which estimate, and confirmations sent by hand. Setup: 2–3 hours — connect the agent, build the nudge automation, turn on booking and reminders, load the wording in the shop's voice. DIY note: this is the plan tier that carries the automations; if the shop is on a lower tier today, the move is a plan change inside a tool it already runs, not a new tool to learn.
QuickBooks Online — Essentials (already in the shop)
What it fixes: L3 and L5. Fix 4's reminder schedule and card-on-file live here, and the Jobber sync in Fix 5 lands its invoices here. What it replaces: chasing invoices from memory, and typing the same job and invoice details a second time. Setup: 1–2 hours — set the reminder schedule, switch on online payments and the card-on-file line, connect the sync, then hand-check the first week. DIY note: every piece of this is a settings screen, not a project.
NiceJob — Starter
What it fixes: the review ask in Fix 3, aimed at the gap named in §6 — 360 jobs finished, 11 reviews received, no record of anyone being asked. What it replaces: remembering to ask, which the records show is not happening. Setup: 1 hour — connect it to the job list, set the ask to go out when a job closes. DIY note: we attach no dollar figure to reviews in this report (§6), so NiceJob is in the stack on effort, not on a promised return: it is cheap, it runs itself, and the ask should never have depended on memory.
CompanyCam — Crew (5 users)
What it fixes: it protects L3. Every job gets time-stamped, organized photos, so when a customer asks "was this finished?" before paying, the answer is a photo with a time on it — not an argument. The same photos can ride along with Fix 2's follow-ups — the agent attaches them to the draft it writes. What it replaces: job photos scattered across four technicians' personal phones, unfindable when they are needed. Setup: 1 hour — app on each tech's phone, one rule: photos on every job, every time. DIY note: honest limit — the Audit measured late payment and write-offs, not disputes specifically. CompanyCam is here as protection for the money leak, and we claim no separate dollar figure for it.
Implementation order
- Jobber first — follow-up automation and reminders on (fixes 2 and 3). This is where the largest stack-addressed dollars sit.
- QuickBooks Online second — reminder schedule, card-on-file, and the sync (fixes 4 and 5), while the Jobber settings are fresh.
- NiceJob third — connect it once jobs are flowing through the new close-out, so review asks start with the next finished jobs.
- CompanyCam fourth — roll it out job by job until photos on every job is just how the crew works.
8Sources appendix
Every figure in this report traces to a record below. Rates used in the math are listed last and labeled as assumptions. Tool prices trace to the vendors' published pricing pages, listed with the other sources.
| Figure used | Source |
|---|---|
| 360 completed jobs; average job value $485; revenue $174,600 | Jobber job report + QuickBooks Online, trailing 90 days ending October 31, 2026 (360 × $485 = $174,600) |
| 412 calls received; 96 missed (23%); 61 missed after-hours/weekend; 35 missed daytime | Call log export, same 90-day window |
| 118 estimates sent; 74 followed up → 30 closed (41%); 44 not followed up → 8 closed (18%); overall close rate 32% | Jobber estimate report, same 90-day window |
| 360 invoices issued; 54 paid >14 days late (15%); average days-to-pay 17; 3 written off = $1,455 | QuickBooks Online invoice aging, same 90-day window |
| 11 new Google reviews in the window | Google review listing, counted for the same 90-day window |
| L1 math: 61 after-hours misses in 90 days ≈ 244/yr; 35% would book ≈ 85 jobs; 85 × $485 = $41,225 | Call log counts, annualized; booking share and job value as below |
| L2 math: 44 unfollowed in 90 days ≈ 176/yr; 23-point gap (41% − 18%) ≈ 40 jobs; 40 × $485 = $19,400 | Jobber estimate report, annualized |
| L3 math: $1,455 write-offs in the window × 4 = $5,820/yr | QuickBooks Online invoice aging |
| L4 math: 4.0 hr/wk × $30/hr × 52 = $6,240 | Office time as reported by the shop; rate as below |
| L5 math: 1.5 hr/wk × $75/hr × 52 = $5,850 | Owner time as reported by the shop; rate as below |
| Jobber Connect (team of 6–10) — ($229 a month billed annually; $299 a month month-to-month) | Jobber published pricing, getjobber.com/pricing — checked November 3, 2026 |
| QuickBooks Online Essentials — ($85 a month at list price; an introductory offer — 50% off the first 3 months — was showing when we checked; the stack math uses list price) | QuickBooks published pricing, quickbooks.intuit.com/pricing — checked November 3, 2026. List price; introductory offers, when shown, are not used in this report's math |
| NiceJob Starter — ($95 a month, no contract; Pro at $135 a month adds AI-drafted replies and referral campaigns — not needed for the review ask itself) | NiceJob published pricing, get.nicejob.com/pricing — checked November 3, 2026 |
| CompanyCam Crew (5 users) — ($119 a month billed annually for Crew with 3 users included, plus 2 more users at $29 a month each; month-to-month runs $149 plus $34 per extra user) | CompanyCam published pricing, companycam.com/pricing — checked November 3, 2026 |
| Assumption — loaded rate, owner time: $75/hr | Stated assumption, used for L4/L5-style time costing and the hours in L1–L3 |
| Assumption — loaded rate, office time: $30/hr | Stated assumption, used for L4 |
| Assumption — 35% of missed after-hours callers would book | Stated assumption, used only in L1; applied to after-hours/weekend misses only |
Tool prices checked November 3, 2026 from each vendor's published pricing — prices change; confirm before buying.
Closing — your options, stated once
This report is yours to keep and use with anyone. The Audit was $999 one time and it is paid — nothing on this page sells it again.
- Do it yourself from this report. The fix order, the do-it-yourself split, and the math are all in here. Fixes 3, 4, and 5 are settings in software you already pay for.
- The Session — $299. Sit down for one hour and set up fix #1 or fix #2 together, on your actual shop — and get coached on working it yourself, so the shop's own people can run it day to day.
- The Retainer — $1,500 / month. We keep this picture current and advise monthly — a refreshed read of your numbers, 3 working calls a month as needed, a straight read before you buy any AI tool, and ongoing coaching on how the owner and the office use AI themselves. Run an idea by us before you spend: bring us a product you're eyeing and we'll check it against your systems — worth it or not — before you buy.
- Your AI Operating Partner (Luke) — $3,000 / month, no initial payment, 3-month minimum. Our agent runs the follow-up and chasing work himself, inside limits you set and with your approval on anything that matters.
Pick any of them, pick none of them, or pick one later. The report works either way.
Full Audit — prepared by AltaPoint Development. SAMPLE — fictional business, illustrative figures. Harborline Plumbing is a fictional sample business; the format, math discipline, and honesty rules are the real ones used in every paid Audit.
After the Audit
Pick any of these, pick none, or pick one later. The report works either way.
- Do it yourself from the report. The fix order and the do-it-yourself split are complete enough to work from alone.
- The Session — $299. Sit down for one hour and set up the first fix together, on your actual shop — and get coached on working it yourself, so the shop’s own people can run it day to day.
- The Retainer — $1,500 / month. We keep this picture current every month: a Monthly Read, 3 working calls a month as needed, straight answers before you buy, and coaching on how you and your office use AI yourselves. Run an idea by us before you spend — bring us a product you’re eyeing and we’ll check it against your systems first.
- Your AI Operating Partner (Luke) — $3,000 / month, no initial payment, 3-month minimum. Our agent runs the follow-up and chasing work himself, inside limits you set.
Questions
Is the Audit a pitch for more work?
No. The report body carries no pitch. Your options are stated once, on a closing page, after the findings. The report is complete enough to hand to any builder.
What if my records are a mess?
Then the report says so. What could not be determined is listed plainly, with what record would settle it. A gap is reported as a gap — never filled in with a guess.
How is this different from the free Snapshot?
The Snapshot is one page built from your answers to five questions, with hours given as a labeled estimate. The Audit measures your actual records and prices each leak in dollars and hours, with sources.
Buy the Full Audit
$999 one time. Written report plus one walkthrough call included.
Want the free picture first? Get a free Snapshot.
HOW WE PRIORITIZE
Every Opportunity, Plotted.
QUICK WINS
High impact, and usually closer than you think — settings, habits, and follow-up you can change first.
MAJOR PROJECTS
Bigger builds, sequenced after the quick wins feed them data.
FILL-INS
Worth doing once the basics run themselves.
IGNORE THESE
What we tell you NOT to build — naming it is part of what you paid for.
The Full Audit plots every leak on this matrix — so you see exactly what to fix, in what order, and what to leave alone.
BUILT FOR CONTRACTORS. ONLY CONTRACTORS.
There isn’t a generic version of this.
AltaPoint is run by a construction project manager here in Long Beach — not an AI agency that added a construction page. Every system is built around how a contracting business actually runs: calls that come in after hours, estimates that sit waiting, invoices that don’t get chased, subs whose paperwork expires.
Start free with the Snapshot — tell us about your business and we will tell you where AI can actually help. Every step stands on its own and is priced on its own: The Session is $299 per session, The Full Audit is $999 one time, The Retainer is $1,500 / month, and the Operating Partner (Luke) is $3,000 / month with no initial payment and a 3-month minimum. If something needs building, it is quoted separately, in writing, case by case.