Handover reports

Proving what was done, by whom, and when.

When a construction company hands a plot back to the developer, it has to evidence the work. That pack protects everyone if a dispute surfaces two years later — and it is usually assembled at the end, from photos in a group chat and somebody's memory. It should be a query, not an archaeology project.

Talk to us See where the records come from

01 — asking for one

One question, one document.

No forms, no export screen, no chasing a foreman for the photos. Ask for a plot and the report comes back — or a plain sentence explaining why it cannot.

Cuprix Handover · Graham Whitmore

Can you get me the handover report for Plot 5 at Bradford?
09:12
Bradford Riverside Apartments, Plot 5 — Riverside apartment block, podium slab. One moment.
09:12
Handover-Report-Bradford-Riverside-Plot-5.docx
6 confirmed work records, signed off by Ian Sagar on 14 Aug 2026.
09:12
And Plot 2?
09:14
Plot 2 hasn’t been signed off yet, so I can’t generate the handover report — the last entry needs to be a foreman sign-off.
09:14

The second answer matters more than the first. A tool that always returns something is a tool nobody can rely on when it counts.

02 — what comes out

A document, not a dashboard.

It has to be filed, emailed and read by people who will never log into anything — so it is an ordinary Word document, laid out the same way every time.

CUPRIX

Handover Report

Bradford Riverside Apartments — Plot 5

Riverside apartment block, podium slab

Generated from confirmed Work Records only. Deterministic query — same plot, same result, every time. Unconfirmed entries excluded from the evidentiary body.

Work Record Trail

Date	Work Type	Description	Operative	Certification	Plant Used
2026-07-28	Setting Out	Podium slab set out to drawing 4102-C, levels checked against site datum.	Wes Duffield	CSCS Gold	—
2026-07-30	Reinforcement	Bottom mat fixed and tied, B785 mesh with T16 edge trimmers, cover checked at 40mm.	Marek Sowa	CSCS Blue · Steelfixer	—
2026-08-03	Formwork	Perimeter shutters erected and propped to podium edge, line and level signed off before pour.	Wes Duffield	CSCS Gold	Telehandler 4T
2026-08-06	Concrete Pour	Podium slab poured, C32/40, 46m³ placed and vibrated. Cube samples taken at 14m³ and 38m³.	Dermot Cassidy	CPCS Concrete Pump	Concrete Pump 32m
2026-08-12	Finishing	Power float completed, surface regularity checked to SR2. Curing membrane applied same day.	Marek Sowa	CSCS Blue · Steelfixer	Power Float
2026-08-14	Sign-off	Podium slab to Plot 5 complete and inspected. Cube results at 28 days to follow under separate cover. Plot released for handover.	Ian Sagar	SMSTS · Site Manager	—

Detail

28 Jul 2026 — Setting Out

Podium slab set out to drawing 4102-C, levels checked against site datum. — carried out by Wes Duffield (CSCS Gold) , verified by Ian Sagar at 2026-07-28 16:42.

Source: Voice note · Confirmed: 2026-07-28 16:42

30 Jul 2026 — Reinforcement

Bottom mat fixed and tied, B785 mesh with T16 edge trimmers, cover checked at 40mm. — carried out by Marek Sowa (CSCS Blue · Steelfixer) , verified by Ian Sagar at 2026-07-30 17:05.

Source: Voice note · Confirmed: 2026-07-30 17:05

3 Aug 2026 — Formwork

Perimeter shutters erected and propped to podium edge, line and level signed off before pour. — carried out by Wes Duffield (CSCS Gold) , using Telehandler 4T, verified by Ian Sagar at 2026-08-03 15:20.

Source: Photo · Confirmed: 2026-08-03 15:20

Photo evidence — Formwork, Plot 5, 3 Aug 2026

…and 3 more records

Sign-Off

Podium slab to Plot 5 complete and inspected. Cube results at 28 days to follow under separate cover. Plot released for handover.

Verified by Ian Sagar (SMSTS · Site Manager) · 14 Aug 2026

Report generated 14 Aug 2026 · Source: confirmed Work Records, joined against roster reference data at generation time · Cuprix

03 — the rules

Six things it will not do.

An evidence pack is only worth having if you can point at it in a meeting and say where every line came from. These are the constraints that make that possible.

01

Confirmed records only

A work record becomes evidence when the foreman confirms it, and not before. Anything still unconfirmed is excluded from the document entirely — it is not shown greyed out or marked provisional, it is simply not part of the evidentiary body.
02

One sign-off, and it must be last

Exactly one sign-off record, and nothing may come after it. Two sign-offs, none at all, or a work record dated after handover was signed — any of those and it refuses to produce a document rather than produce a misleading one.
03

Names are stored on the record

Who did the work, what they were certified to do and what plant they used are written onto the record when it is confirmed. Someone leaving the company a year later cannot quietly change what a finished report says about who was standing there.
04

A photo is either evidence or it is absent

If a record has a photo, the report embeds it. If it references one that cannot be found, that is flagged as a broken reference. If it has none, no image appears — never a generic placeholder, because a stock graphic in an evidence pack looks like proof of something and is proof of nothing.
05

The content is never cached

The list of sites and plots is cached, because it barely changes and re-reading it on every run is waste. The work records themselves are always read live. That distinction is the whole guarantee: a report is what the record said at the moment it was asked for.
06

Same plot in, same document out

One parameter, one fixed query, one fixed layout. There is no step where anything decides what is relevant or how to phrase it — which is what makes two people running it a month apart get the same document.

04 — when it can't

Every way it can fail has its own answer.

Not one vague “couldn’t generate it”. You still need to know which of these happened — just not in technical terms.

Configuration error Something is set up wrong at our end — not something to retry blindly.
Cannot reach the records “I couldn’t reach the system right now — try again in a moment.”
Plot not found “I couldn’t find a plot called that — want to pick from the list instead?”
No confirmed work “There’s no confirmed work logged for that plot yet, so there’s nothing to put in a report.”
Integrity failure “This plot hasn’t been signed off yet, so I can’t generate the handover report.”
Render error It did not work, and making something up instead is not an option.

Each one is a distinct signal, so an automated trigger can branch on the reason without reading a message — and so nothing downstream ever mistakes “no records yet” for “nothing happened”.

05 — two ways to run it

A person asking, or nobody at all.

Someone asks

In a chat, in plain English

A director asks for a plot and gets the file back in the same reply. If the plot is ambiguous it asks which site first, then which plot. No terminal, no install, nothing to learn.

Or nobody does

Fired by something else

The same logic runs as a single command, so a workflow, a scheduled job or a message to the site number can produce the report with no one in the loop — and can tell the difference between the failures above.

Both paths run identical logic. A report generated at 3am by a scheduled job is the same document, to the character, as one asked for in a chat.

06 — recorded

Where this one came from.

Recorded · Instagram Turning a client's problem into something real Proving exactly what work was done, by whom, and when — pulled straight from the records, in seconds.

Watch this on Instagram

07 — from here

Two ways to go from here.

See the records

Where the evidence comes from

A report is only as good as what feeds it. The workspace shows the captures behind the records — the actual words a foreman spoke, and what they became.

Open the workspace

Or bring your own

It works on what you already have

This is not tied to one system. Whatever you already use to log site records, the same approach sits on top of it — the hard part is agreeing what counts as confirmed, not the document.

Get in touch