Solutions plan — every open problem, step by step

Written 2026-09-15, on Omar's order ("go through every scenario, step one, step two, step three — I need a solution for all the problems; money is not an issue").

Source for the problems: the Fieldflo call notes (Fieldflo_Competitor_Call_2026-09-08.md), the capabilities list (App_Capabilities_2026-09-15.md), and Omar's answers on 2026-09-15. All app data is demo until Omar says otherwise. Nothing in here sends anything to anyone.

Closed, do not reopen: five companies in one login. Omar, 2026-09-15 20:59 — Makki Demolition only.

1. Labourers answer by text, no login

The problem. the owner liked that his men never have to open an app. Ours makes a labourer log in. Men don't log in.

The solution. One Makki phone number. Two texts a day. The man replies with one character. No link, no login, no app.

1. Buy a Canadian phone number (Twilio, ~$1.50/month + about a cent a message). Roughly $10–20/month for a crew of 20.
2. The owner sends the crew list: name + mobile. Only numbers on that list are ever answered — anything from an unknown number is ignored and shown to the foreman.
3. Morning text, 6:45 a.m., to every man scheduled that day: "1425 17 Ave SW. Toolbox talk: silica. Reply Y to sign off." Y → his name and the time land on that job's safety record. That is the signature an inspector asks for.
4. End-of-day text, 4:30 p.m.: "Hours today: 8.5. Reply Y if right, or reply the number." Y → confirmed. 8 → logs 8 and flags it corrected so the foreman sees the change.
5. Anything that isn't Y or a number gets one plain re-ask. Still nothing → it goes on the foreman's list, and the foreman phones him. No loops, no nagging.
6. No reply by 5:30 p.m. → the hours stay unconfirmed and the owner's dashboard shows it, same as today.
7. The Crew login stays for the men who like it. Nobody is forced either way.

Needs from Owner: the go-ahead to buy the number, the crew phone list, and a yes on the two wordings. Status: planned on paper (Labourer_By_Text_Plan_2026-09-15.md). Not built. Nothing sends.

2. The owner runs the whole business by text

The problem. the owner's #1 ask from day one: text it and it gets done. Today he has one Telegram line and four helpers behind it.

The solution. The same number as §1 handles the owner, and knows it's him.

1. The owner texts a question — "what's outstanding", "who's on 17 Ave", "send me the Monday tickets" — and gets the number straight back, from the app's own data.
2. The owner texts an instruction — "bump Crew B to Thursday", "mark 17 Ave complete" — it makes the change in the app and texts back what changed, in one line.
3. Anything that costs money or leaves the building — an invoice, a quote to a customer, a text to a supplier — is written up and handed to the owner. He says send. It never sends on its own.
4. Foremen keep sending voice notes; those get turned into the daily report the same way they are today.

Needs from Owner: nothing new. This runs on what's already built. Status: working today on Telegram. Moving it to a plain phone number is part of §1's number.

3. QuickBooks

The problem. the owner runs on QuickBooks Online. Ours is a preview screen showing what would sync. Until it's real, he's still typing things twice.

The solution.

1. The owner clicks one "Connect QuickBooks" button and signs in with his own QuickBooks login. We never see his password.
2. First pull, read-only: customers, invoices, payments, chart of accounts. He looks at it and says whether the numbers match his books.
3. Only when he says they match do we turn on the push: an approved invoice in the app becomes a QuickBooks invoice. One button, his click.
4. After that it checks every night: payments that landed in QuickBooks mark the job paid here, so collections and outstanding are always the real figure.
5. Payroll hours from §1 come across as a timesheet export, not a direct write to payroll. Payroll stays where the owner runs it now until he says otherwise.

Needs from Owner: his QuickBooks login, at his screen, one time. Fifteen minutes. Cost: nothing extra. Uses his existing QuickBooks plan. Status: preview screen built. Real connection not started.

4. Asbestos and abatement paperwork

The problem. This is the part of Makki's work nobody else's software covers, and the part that costs money when it's wrong. Right now we have a document shelf you can upload to — that's all.

The solution. Build the four documents the regulator and the insurer actually ask for, off the data the crew is already entering.

1. Containment log. Per abatement job: date, area, containment type, negative-air readings, smoke test, who was in the enclosure and for how long. The crew fills it on the Start Day screen; the men's names come from the crew list; times come from the hours in §1.
2. Waste manifest. Per load out: generator, hauler, landfill, material, weight, ticket number, photo of the scale ticket. Prints as the manifest, and the copies stay attached to the job forever.
3. Regulator / OHS report. One button on the job: prints the file an inspector asks for — tickets and expiry for every man who was on site, toolbox talks signed, containment log, waste manifests, photos, incident notes. In order, dated.
4. Claim walkthrough. For insurance work: a room-by-room record with photos before, during and after, scope per room, and the same sheet exported for the adjuster.
5. Everything above is already half-there — photos, tickets, talks, cost codes. This turns loose pieces into the four documents.

Needs from Owner: copies of the forms he uses today — his containment log, a waste manifest, whatever he sends the adjuster — so ours look like his, not like a guess. Status: not built. Biggest remaining piece of real work.

5. Fifteen tabs down to seven

The problem. Owner login had 15 tabs. Too many doors to the same rooms.

The solution. Seven groups: Dashboard · Jobs · Schedule · Money · People · Safety · Gear. Nothing cut, every screen still reachable, the other four logins untouched.

Status: building now. Reports back with pictures.

6. The slides

The problem. Slides 6–8 used cropped pieces of screens. Rejected. Omar wants one workflow per slide and one whole page.

The solution.

1. Agree the three workflows first. Recommended: (a) quote to crew on site, (b) a day on site, (c) job done to money in the bank.
2. Per slide: left side the numbered steps in plain words, right side one full screenshot of the single page that carries that workflow, numbers dropped on that page.
3. Omar approves the plan on paper before anything is rendered.
4. Then rebuild, one slide at a time, good / bad / change on each.

Status: frozen until step 1 is agreed.

7. Real data instead of demo data

The problem. Every figure in the app is invented for the demo. It looks right, it isn't his.

The solution, in the order it should happen.

1. QuickBooks connects (§3) — that brings customers, invoices and payments across on its own.
2. The owner sends the crew list: names, phones, ticket types and expiry dates. That makes tickets, crews and the blocking rule real.
3. The owner sends the equipment list. That makes Gear real.
4. The owner sends a handful of past quotes. That's what the Estimator learns from.
5. Job addresses become real map pins instead of demo placements.
6. On the day we switch, the demo data comes out in one move and the DEMO header comes off.

Needs from Owner: four lists and one login. Nothing technical. Status: waiting on the owner.

8. Putting it live

The problem. Everything built since this afternoon sits on this machine. The link the owner has still shows the older build. Also: the live link is on EazyLife's Cloudflare, not Makki's.

The solution.

1. Omar says "go" per push; the build goes live and the link is re-checked.
2. Decide whose account hosts it long term — The owner's own Cloudflare, or EazyLife's. The owner's is the cleaner answer: his app, his account, nothing to untangle later.
3. When the owner is ready, point a Makki address at it (app.makkiinc.ca or similar) instead of the pages.dev link.

Needs from Omar: the "go", and the hosting decision.

9. The four helpers not built yet

Each one is a day or two of work and each is blocked on one thing from the owner.

The order I recommend

1. Now: seven tabs finishes (§5). Agree the three workflows (§6).
2. This week: buy the text number and build the labourer texting (§1, §2). It's the thing the owner said out loud that we don't have.
3. Fifteen minutes with Owner: QuickBooks (§3). Everything downstream gets easier once it's real.
4. Then: the abatement paperwork (§4). That's the moat — nobody else builds it for asbestos.
5. Then: real data (§7), hosting (§8), the four helpers (§9).

What it costs

• Text number: about $10–20 a month, all in.
• QuickBooks: nothing extra.
• Hosting: free at this size.
• Everything else is build time, inside the price already agreed ($3,000/month for the build, then $500/month upkeep).