Bring this to the operations sit-down and take notes directly into it (the checkboxes and blanks are live on screen). Goal: leave with enough to lock the data model, the RBAC matrix, and the MVP scope. Feeds BUILD_PLAN.md Part 1 (Discovery) and Part 4 (Data Model).
Tip · Print or "Save as PDF" → it switches to a clean, write-on-it layout
Talk to one real person per role — not the owner guessing what the field crew does. Get an actual party chief, the person who runs payroll, the office scheduler.
"Show me, don't tell me." Watch them do the task and see the actual files, spreadsheets, and screens. The real workflow lives in the workarounds.
Capture their vocabulary verbatim. Their words for statuses, job types, and fields become your schema — don't translate them.
Quantify everything. "A lot" is useless; "~25 jobs/week, payroll takes 4 hours" is a baseline you can prove ROI against.
Listen for exceptions. "Usually X… but sometimes Y" — the Y is where software breaks. Write down every "sometimes."
Don't pitch features. Capture the need and the pain; solutions come after.
Record it (with permission) or bring a second note-taker.
Listen hard for hidden work / red flags: "we just know" (tribal knowledge), shadow spreadsheets nobody mentioned, manual re-keying of the same data into two systems, and "I text the crew" / "I call to check" (the communication gaps you're replacing).
0
Company-wide framing
Ask everyone; it sets context and surfaces the wedge.
What does a typical week look like? Seasonality / busy vs slow? __________
The single thing that wastes the most time today?__________
The single thing that, if it broke, would hurt the most? (lost packet, missed deadline, lapsed license, payroll error) __________
If we fixed one thing in the first 6–8 weeks, what should it be?__________
1
Trace a real job end-to-end (the spine)
Pick 2–3 recent, representative jobs and walk each through every stage. At each step capture: who does it, what fields, what tool / where it lives, how long, and the handoff. This becomes your data model.
Job number scheme — exact format, who assigns, sequential / per-year / per-client? __________
Exact status values & order (storyboard assumes Assigned → In Progress → Waiting → Completed — is that real? what triggers "Waiting"?) __________
What "done" means and who decides a job is ready to invoice __________
Where the current packet comes from (Ignite? email? Drive?) and how a revision reaches the crew __________
2
Role-by-role question banks
Company Owner
How do you know today whether a job was profitable? ______
What numbers do you wish you saw daily / weekly? ______
Revenue & margin by job type — which work is profitable vs not? ______
Growth / headcount plans the platform should support? ______
Who may see pay, financials, profitability? ______
Project Manager
How do jobs come in (call, email, repeat, bid)? Who creates them? ______
How do you decide which crew gets a job? Constraints — skills, certifications, equipment, location? ______
How do you track status now, and answer "what's the status of X"? ______
What documents exist per job and who produces them (CAD/drafting)? ______
How do crews learn a packet was revised? ______
Which handoffs break or get dropped? ______
Office Staff / Scheduler
Walk me through building the schedule — how far ahead, what tool? ______
Timesheet collection: how, how often, what format? ______
Payroll prep: every step, how long, what tool. Pay types (hourly/salary/1099), OT rules, per-diem, mileage, bonuses? ______
Invoicing: who, when, what system, how is AR / unpaid tracked? ______
Where do customer records live? ______
PTO requests & approvals — current process? ______
Field Crew (talk to an actual party chief)
What phone/tablet do you use? Carrier? How's signal — how often no service? ______
How do you get today's assignment + the packet right now? What's annoying? ______
What do you need on site (maps, plats, control points, deed, instructions)? ______
How & when do you report back — notes, photos, issues? On site or end of day? ______
How do you track time / mileage today? ______
What makes you stop and call the office? ______
HR / Payroll (often the owner or office — confirm who)
Certifications/licenses tracked (PLS, CST, OSHA-10, CDL)? Who owns renewals, how reminded? ______
How are pay rates kept confidential today? ______
Tax forms (W-2/1099) — stored where, or handled by a payroll provider? ______
New-hire onboarding steps? ______
3
Tools being replaced (migration inventory)
For each tool: what it does, who uses it, the data in it, must-keep features, cost, export format. The cost column is also your consolidation sales pitch.
Tool
Purpose / who uses
Data held
Must-keep / export
Cost
Ignite (job distribution)
Spreadsheets
Google Drive / file storage
Accounting (QuickBooks/Xero?)
Payroll tool
PTO tracker
Texting / group chats
_other_
4
Data & documents
Document types per job: PDF, CAD (DWG/DGN?), plats, deeds, permits, photos, point/coordinate files (.csv/.txt/ASCII), field books ______
Rough volumes: documents per job __ · total storage today __ GB · largest file types ______
Naming conventions for files & jobs ______
How long must records be retained (legal/contractual)? ______
Capture today's numbers so you can measure improvement after launch.
Office-to-field calls/texts per day: ______
Payroll prep time per cycle: ______
Time to find a historical job or document: ______
How often a crew runs the wrong/old packet: ______
Average job turnaround (intake → delivery): ______
Any current app — % of crew using it daily: ______
9
MVP scope (freeze it before building)
What must Phase 1 include for the team to actually adopt it? ______
What would make crews open the app daily? ______
What is explicitly out of scope for v1? ______
Who are the pilot users (one crew + one PM + the owner)? ______
★
After the meeting — produce these
Workflow diagram — the job lifecycle from §1 (becomes the dispatch/status logic)
Data dictionary — every entity, field, and exact status value (→ Part 4 + first migrations)
Tool migration inventory — §3, with what data must migrate
RBAC matrix — for each role × each sensitive resource: allowed / denied (your most important test suite)
MVP scope statement — §9, written down and protected
Success-metric baselines — §8 numbers, to measure against post-launch
Bottom line: the parts that change based on this meeting are mostly the job fields, statuses, and role permissions — everything else (auth, app shell, design system, infra) you can already build. Capture §1 and the RBAC matrix well and you can start coding the domain the next day.