Master Scan · Dispatch pilot

Tomorrow's roster, ready by 8pm

Every request read and checked against your rules, the whole day planned together. Your supervisors look it over, change what they want, and send.

Tonight, as it is

Built by hand, every evening. WhatsApps, calls, passes, leave, calibration dates: mostly from memory.

With it

Waiting for you at 8pm. On a made-up day: 17 jobs ready to approve, 6 to look at, 1 flagged with the reason.

Three questions for today
  1. What would your evenings look like with it?
  2. How would it work, and how is it built?
  3. What would we do together, week by week?

Every screen in this presentation uses made-up data: names, sites and requests are all invented.

01
What I heard

Today, the whole day rests on one experienced supervisor

about 1001field staff
about 30 a day1requests, mostly the day before
3 waysWhatsApp, calls, email, all free text
About 4supervisors match by hand, led by one with 15 years

It works because he holds it all in his head. The aim is to write that down, so it is there on his days off too.

Where it hurts, in the order you ranked it
Version 1 goes after the first three.
1Wrong person, permit or levelA replacement, a complaint, lost trust
2Missing kit or checklistHours lost going back; a safety risk on oil and gas sites
3Supervisor timeFour people matching by hand every evening
4Room to growMore work without hiring
5Van routesSix vans, instructions on the day
6Visibility and billingSmaller sites unseen; invoices matched by hand
Your figures, from our conversation on 1 October. Escalations suggest about 5% of jobs hit a problem, and you think it's higher.
02
The idea

The system recommends. Your supervisors approve.

1Requests come inWhatsApp, email, calls. One new habit: a supervisor forwards each request, one tap.Your team
2Each becomes a job cardSite, date, method, level, people, conditions. Anything unclear is flagged.System
3Matched and checkedOnly people who pass every rule. A kit list each, calibration checked.System
4Plan ready by 8pmTomorrow's whole roster on one screen.System
5Approve or changeThen each person gets their job and kit list on WhatsApp.Supervisor
The system does it   A person decides
In version 1
  • Reading every request into a job card
  • Matching on the rules, then ranking
  • A kit list for each person, calibration checked
  • The 8pm review: approve, or change with a reason
  • A WhatsApp message to each technician, sent by the supervisor
Later, only if version 1 earns it
  • A phone app for technicians
  • Van routes
  • An owner's view of the whole team (appendix A1)
  • How much more work the team can take
  • Billing reconciliation
03
One request, start to finish

A WhatsApp at 3:42pm becomes a checked job by 8pm

One request's journey
A made-up Wednesday 7 October, planning Thursday 8 October.
WhatsAppEmailCall
It arrivesHowever it comes in
It's readInto a job card
Thu 8 Oct
Tank terminal · UT L2 × 2
In the planNext to every other job
Prakash D.
Hari C.
Matched18 of 100 pass every rule
KittedEvery unit in calibration
✓
Approved and sentBy the supervisor
Tank terminal · site repclient
3:42 pm

It arrives the way it does today.

Shorthand is normal: "tmr", "Lvl 2", "same as last week". Your clients change nothing.

Each coloured phrase lands in a field. Anything the message doesn't say is marked for a check.

Who can go: every rule, one at a time

Each person gets their own kit list, and nothing out of calibration goes out.

"Calibrated until" is the date each unit's calibration certificate runs out. If it runs out before the job ends, the unit is swapped for one from the store.

Master Scan Opssupervisor
OK 👍

Prepared by the system. Sent by the supervisor.

Site, gate, report time, partner, and the actual units with their dates. No new app.

Made-up data, run through the working demo's own rules and records.
04
The whole day at once

At 8pm, one screen: 31 jobs, 57 people, planned together

Thursday 8 October, as the supervisor sees it at 8pm
24 new requests plus 7 multi-day jobs already running. Planned in under a second.
JobNeedRecommendedStatus
Six of the 31 rows, from the same made-up day.
The one it can't fill, and says so

Refinery A, Jurong Island: radiography, a crew of 2. Only one licensed radiographer is free. Nobody goes out alone on radiography, so it flags the job: call the client, move it to Friday?

Try it: change a pick on the tank terminal job
Pick a reason, then a replacement. Anyone you can pick already passes every rule; blocked people are greyed out.
05
How it's built

This is real custom software, and it fits around the tools you already use

Three layers
Requests come up from your tools, the engine plans, your people decide.
WhatsAppforwardedEmailand PDFsCallstyped or saidYour sheetsstaff, kit1 Your tools, as today ReaderAI readsRulespasses, certsPlannerwhole dayKit checkcalibration2 The engine 8pm reviewapprove or changeTechniciansjob and kit list3 Your people Evening summaryto you and meRuns in your owncloud account,in Singapore
Drag to turn it.
How a request flows through it
Green is working in the demo today, orange was tested on messy requests, grey is what I build for your version.
Working in the demo Tested on 14 Oct To build WhatsApp forward to one dispatch number Email dispatch inbox, PDFs too Phone call note typed or dictated Paste box anything else Cloud account in Master Scan's name, Singapore One inbox every request, with where it came from and when Reader: Claude, through the API text into a job card, against a fixed form; anything unclear is flagged, never guessed 198 of 202fields right on 15messy requests Your sheets staff, sites, kit, read each evening Rules level, pass, certificates Planner OR-Tools, whole day Kit check every unit's calibration 8pm review screen approve, or change with a reason; phone or laptop, signed in The demo has aclickable version Log of every event plans, approvals, changes and reasons Health checkto us, counts only WhatsApp to each technician prepared by the system, sent by the supervisor with one tap Job and kit list on the phone they already use. No app to install in version 1.

Four ways a request comes in

01
WhatsApp

A supervisor forwards the client's message to one dispatch number. One tap.

02
Email

Sent or forwarded to a dispatch address, PDF work orders included.

03
Phone calls

After a call, the supervisor types or says a line.

04
Anything else

A paste box for the odd one out.

Each piece, and what to check with you

Only the last column needs you: tell me where today is different.

PieceWhat it doesWhat it usesPlease check
WhatsApp inClient requests reach the systemA dispatch number on WhatsApp's official business serviceDo clients message a business number, or supervisors' personal phones and groups? Either works: personal chats are forwarded
Email inRequests and PDF work ordersOne dispatch inboxMicrosoft or Google email? Any client portals?
Calls inA note after each callA "new request" box; type or speakRoughly how many requests come by phone?
Your recordsStaff, sites, equipmentYour own Excel or Google Sheets, read each eveningWhere do they live, and who keeps them up to date?
ReaderMessage into a job cardClaude, Anthropic's AI model, an outside service reached over the internetAny client whose contract limits outside services?
RulesWho may go whereWritten from your sheets and your senior supervisorWhich rules live only in people's heads?
Day plannerThe whole day at onceGoogle's free planning software, OR-ToolsNothing; this part is mine
Kit checkEvery unit in calibrationYour equipment registerIs calibration tracked per unit and probe today?
The 8pm screenApprove or changeA web page; phone or laptopDo supervisors plan on a phone or a laptop?
Messages outEach technician's job and kitPrepared, then sent from the supervisor's WhatsAppDo technicians all use WhatsApp?
Where it runsThe app and its recordsA small cloud account in Master Scan's name, in SingaporeMicrosoft or Google account preferred?
Evening summaryDid it run, what it planned, anything failingA short report to you and meWho else should get it?

What goes into it

Already working
2,138
lines of code in the demo, with 16 automated tests
  • The rules engine
  • The whole-day planner
  • Kit and calibration checks
  • The replay test
Built for your version
Reading four intake routes
Connecting your sheets and register
Your rules, from your senior supervisor
The 8pm screen, made good to use
Technician messages
Hosting in your account, sign-in, the log
The evening summary
Testing on your past requests
Bar length is my rough estimate of effort.
2,400person-and-job pairs checked every evening
5kinds of check for each pair
636equipment units with calibration dates
0.06sto plan the whole day
On the made-up day: 100 staff and 24 new requests.
06
Safeguards

Nobody goes out without a person's approval

01It never dispatches anyone on its own

A supervisor approves every assignment and sends every message.

02A gap takes that person off the list

A missing pass, an expired certificate or an out-of-date unit takes that person or item off the list.

03Changes need a reason

Swap in anyone who passes the rules. Overriding a block is a separate step that records who and why.

04Everything is logged

Every plan, approval and change: who, what, when.

05We both see it every evening

A short summary to you and me: it ran, what it planned, anything failing. Counts only.

07
Proving it

Before anything goes live, it is tested on your own past weeks

1Count today

Three numbers a day, a few minutes, on a sheet I set up.

2Replay

Your last 2 to 4 weeks run through the system; differences go to your senior supervisor.

3Compare

Weeks with it against weeks without.

1 · What we count, every day
Wrong allocationsA replacement sent, or a client unhappy with who came0 · 1 · 0 · 2 · …
Missing-kit incidentsSomeone went back, or waited, for equipment or paperwork1 · 0 · 0 · 1 · …
Minutes allocatingTime the supervisors spend building tomorrow's plan95 · 110 · 80 · …
Example entries only. If you go ahead, this is the one thing worth starting early: a week that isn't counted can't be counted later.
2 · The replay: each column is a past day
20 made-up days, 576 jobs. Did the system pick the same people your team sent?
Same peopleSome of the sameDifferent: a card to review
Made-up example · Tue 15 Sep · Tank terminal, Jurong Island · UT L2 × 1

The team sent T-021. The system picked T-033 (UT Level 2, in the site's usual team).

Would that have been right?

YesNo: needed restNo: client preferenceNo: other

3 · How we'll know it worked (the KPIs)

MeasureCounted byGood looks like
Wrong allocations per weekYour tallyDown
Supervisor minutes a day on allocationYour tallyDown
Missing-kit incidentsYour tallyDown
Plans approved without changesThe system's logUp, as your reasons become rules
08
Working together, week by week

You're my co-pilot: each week you get something to react to

Before we start
You getThe count sheet and a one-page note on how your data is handled
Your inputStart counting; share the four things
Week 1
You getThe system running on your sheets, and your rules written down for you to check
Your input30 minutes with your senior supervisor
Week 2
You getThe replay report on your past weeks, every difference explained
Your inputAbout an hour on the difference cards
Week 3
You getThe system's plan at 8pm every evening, next to your supervisors' own, and the evening summary (email and the paste box first if the WhatsApp number isn't ready)
Your inputSupervisors keep leading as today; the system runs alongside and offers alternatives
Week 4
You getImprovements from your supervisors' reasons, and the counts side by side
Your inputWhat's annoying, what's missing. When you're comfortable, supervisors start from the system's plan
End of week 4
You getA one-page result
TogetherKeep, change or stop
Your data
  • Your people by name on your screens; the AI step works with IDs.
  • Client and site names are swapped for codes before the AI step wherever we know them, so it sees as little as possible.
  • It stays in your account, used only for this, deleted whenever you ask.
  • The AI service keeps nothing for long: deleted within 30 days2 and not used for training without permission3; we never give that permission.
  • A one-page note on all of this before you share anything.
How we'll work
  • Free. I'm doing this to take load off your supervisors, and I'll learn from it too.
  • Runs on your tools. If it ever stops, your team carries on as today.
  • We both watch it. An evening summary goes to you and me.
  • You own it. Your setup and data are yours.
  • Either of us can stop at any time.
09
Next steps

Starting is small: counting can begin this week

  1. Start three simple counts. Wrong allocations, missing-kit incidents, minutes spent allocating. I'll set up the sheet.
  2. Share four things once you're happy with the data note: staff skills sheet, sites and what each requires, equipment register, two to four weeks of past requests with who was sent.
  3. Thirty minutes with your senior supervisor in week 1.

A few questions for you

Your answers decide how version 1 is wired. Notes stay on this laptop.

YouStart the three counts. I'll send the sheet.
MeSend the one-page data note.
YouShare the four items once you're happy with the note.
TogetherPick a time for 30 minutes with your senior supervisor.
10
After we met, 23 Oct

What you told me, and what happens next

What you told me

  1. 30 to 50 requests a day, 50 at the most, with no cut-off time. Most come to the WhatsApp Business number; email runs on Google Workspace; one client started sending through its own portal in the last 10 days. Calls go into a notebook, then into Excel.
  2. Three departments, and one supervisor handles 80 to 90% of the work. That is where the system helps most, and the risk when he is away.
  3. The rules live with that supervisor today. They are easy to write down, and your supervisors will help.
  4. Permits are set per project, not per client: two lines at the same site can need different permits. Expiries are tracked in Excel.
  5. Calibration is checked by hand against the equipment list, because the automatic reminders did not stick.
  6. Supervisors plan on a desktop, with the phone as a second option.
  7. The tools you have been shown needed constant data entry and never recommended who to send.

The first month, as we agreed

  1. Two weeks testing on your past requests, comparing the system's picks with who was actually sent.
  2. Then two weeks with your supervisors leading as today, and the system alongside offering alternatives. It moves to recommending first only when you are comfortable.
  3. Every week you get the three counts: wrong allocations, missing-kit incidents and minutes spent allocating.
  4. No fixed deadline. Your data stays in your own Google cloud account in Singapore, and client and site names are coded before the AI step.

Next steps

YouSend me the breakdown of who handles which requests across the three departments.
YouShare this with your team, and set up 15 to 20 minutes for me with your software engineer.
MeSend the count sheet and the one-page note on how your data is handled.
TogetherStart the three counts, and meet again to go through the details week by week.
A
Other ways to see it

Ideas for later, for you and your supervisors

Illustrations on made-up data, to spark ideas. I'd build any of these only if you want them.

A1 · Your view as the owner

A2 · The supervisor's phone at 8pm

Approve the ready jobs once you've looked them over. Tap an amber job to see why it needs a look; tap red to see who's short.

A3 · The plan as a board

Tap a person and only the jobs they're cleared for light up; tap a job to see who can take it and why the rest can't.

  1. Your figures Your own figures, from our conversation on 1 October (what you told me; not checked against your records)
  2. Confirmed Anthropic automatically deletes API inputs and outputs within 30 days of receipt or generation by default. (privacy.anthropic.com)
  3. Confirmed Data Anthropic retains from the API is not used for model training without the customer's express permission. (platform.claude.com)
Dispatch pilot for Master Scan · prepared by Ajay · October 2026