The asymmetry nobody mentions when you buy the policy
On one side of a claim is an organisation that handles thousands of them, with systems, checklists and people whose entire job this is. On the other side is you, handling one, for the first time, in the weeks when the business can least afford your attention.
That asymmetry is not unfair — it is just structural. But it does mean the burden of assembling evidence falls on the party least equipped to carry it, and it falls at the worst possible moment.
The consolation is that the work is more mechanical than it looks. A claim file is not a piece of persuasion; it is a schedule of documents that adds up and can be checked. Most of what makes it hard is volume and format, and both of those are solvable problems.
This page is about the practical shape of that work: what the pack contains, where the hours actually go, what doing it badly costs, and — just as importantly — where a document reading tool stops being useful.
What we do not do
Worth being blunt about this before anything else, because the claims market is full of services that blur the line.
We do not assess or value claims
What your loss is worth under your policy is an assessment. That is the job of a loss assessor acting for you, your broker, or your solicitor.
We do not negotiate or submit
Nothing is sent to an insurer on your behalf. The schedule and the documents are yours to present.
We do not advise on cover
Whether your policy responds, and on what basis, is a question about wording. It is not a question about documents.
We do not verify that documents are true
Reading a document tells you what it says. Whether what it says is accurate is a different question, and an important one in claims.
What is left is the part that consumes the most hours and the least judgement: getting a few hundred documents out of PDFs, photographs and envelopes and into a table that sums correctly.
What an evidence pack actually is
People imagine something formal. It is three things, and only the first one takes real work.
| Component | What it is | Why it matters |
|---|---|---|
| The schedule | One row per document, with amounts and a source reference | It is what gets checked, and everything else supports it |
| The documents | Supplied in schedule order, not the order they were found | Order alone can save a fortnight of back-and-forth |
| The covering note | One or two pages: what happened, how figures were built, what is estimated | Turns discoveries into disclosures |
Notice what is not on the list: a narrative of how difficult the period was, photographs of everything at maximum resolution, or every email exchanged with a supplier. Volume works against you, because a file too large to read gets sampled — and sampling is where inconsistencies are found.
The covering note is the cheapest component and the most consistently skipped. Two pages saying what was estimated and why converts your weakest figures from things to be discovered into things you disclosed.
Where the time actually goes
People budget for the wrong parts. The thinking — what to claim, how to present it — is a couple of hours. The typing is days.
| Task | By hand | With extraction |
|---|---|---|
| 200 receipts into rows | 6–8 hours, with errors | Minutes, then a targeted review |
| A year of bank statements | A full day of retyping | One pass, all months in one table |
| Finding a document behind a figure | 10–20 minutes each time | Seconds, from the row reference |
| Checking the arithmetic | Rarely done at all | Automatic, discrepancies shown |
| Splitting net, VAT and gross | Often skipped, then redone | Separate columns from the start |
| Rebuilding after a question | Hours, and the total moves | Filter and re-export |
The last row is the one that decides how a claim feels. Questions always come, and a file that can answer them by filtering a table is a twenty-minute job each time. A file held together by manual typing has to be partly rebuilt, and every rebuild risks the total moving — which is the single most damaging thing that can happen to a claim file.
What doing it badly actually costs
The obvious cost is time, and it is the smallest of the three.
Under-claiming. The most common and least visible loss. Small increased costs of working — a hire charge, expedited freight, a fortnight of overtime — are individually trivial and collectively significant, and they are exactly what gets dropped when someone is retyping receipts at eleven at night. Nobody ever finds out what was left on the table.
Elapsed time. Each round of questions costs a fortnight of calendar time on both sides. Four rounds is two months, during which the person running the business is thinking about the claim instead of the business. That cost never shows up in a settlement figure and is often the largest one.
Credibility. Arithmetic that does not tie, figures that change between versions, a gap someone else finds — each converts a review of the numbers into a review of how the file was assembled. That is a much slower conversation, and it applies to figures that were never in doubt.
A claim with clean records and a checkable schedule: front-loaded effort, one or two rounds of questions, finished.
A claim assembled as questions arrive: effort spread over months, four or more rounds, and the smaller costs never claimed at all.
The same total effort produces those two outcomes. The only difference is whether it happens in one concentrated block at the start or in fragments across a year.
The workflow, end to end
1 · Capture what perishes
Photographs before clean-up, thermal receipts scanned, exports taken from any system that might be replaced.
2 · Get everything into one place
One folder, whatever the format. Sorting comes later; gathering has to come first.
3 · Read it into rows
Up to 100 files per pass — invoices, receipts, statements — into one table with the source file and page.
4 · Check the arithmetic
Lines to totals, VAT to rate, schedule totals to the accounts they should reconcile to.
5 · Categorise
Revenue, continuing cost, saved cost, increased cost of working. Four categories carry most claims.
6 · Write the note
What happened, how the figures were built, what is estimated, what is missing. Two pages.
Steps one and two are the ones with a deadline attached. Everything after them can be done next month; those two cannot, because what they capture stops existing.
A realistic first week
Assuming the loss was recent and you are doing this yourself, alongside everything else.
Day one, two hours. Photograph everything before anything is cleared. Scan the receipts that are already fading. Export from the till or booking system. This is the only day where delay costs you evidence permanently.
Day two, one hour. Find the policy, write down the four dates that matter, and email your broker to confirm the indemnity period. The answer will take a few days, which is why the question goes out early.
Day three, three hours. Gather. Bank statements for the period and the two prior years, invoices, the receipt box, payroll. Not sorted — gathered. One folder.
Day four, two hours. Read it all into rows and look at what came out. This is the point where the shape of the claim becomes visible, and usually where you discover two things you had forgotten about.
Day five, two hours. Categorise, run the checks, and write the list of what is missing. That list is what you spend the following fortnight chasing, while the documents still exist to be chased.
Ten hours in the first week, and the claim is in a shape where somebody else can help with it. The alternative — starting when the first request arrives — is the same ten hours spread over three months, done in fragments, with the perishable evidence already gone.
Four scenarios
A flooded café.Eleven weeks closed, two hundred receipts for the reinstatement, three years of till data for the baseline. The heavy lifting is receipts and prior-year takings, both of which are pure volume. The judgement — how much of the second site’s uplift to deduct — takes twenty minutes once the numbers exist.
A workshop fire. Equipment destroyed, a subcontractor engaged to keep orders moving. The claim turns on purchase invoices for the equipment, most of them years old and buried in an accounting system nobody has opened since the last accountant left, plus the subcontractor invoices that arrived during the disruption.
A theft from a van. Small in value and disproportionate in effort, because it is entirely receipts — some on the phone, some in a glovebox, several already illegible. Speed matters here more than anywhere, because thermal paper does not wait for anyone.
A supplier failure interrupting production. No physical damage at all, so the whole claim is documentary: orders that could not be fulfilled, expedited alternatives sourced at a premium, and the price difference evidenced line by line. This is the type that most rewards good records and most punishes poor ones.
Four very different events, one identical bottleneck: a few hundred documents in the wrong format, standing between you and a schedule that adds up.
Who does what
| Role | What they are for | Who pays |
|---|---|---|
| You | Gathering documents and knowing what happened | — |
| Broker | The wording, the notification, the relationship | Usually via commission |
| Loss adjuster | Establishing the facts for the insurer | The insurer |
| Loss assessor | Preparing and arguing the claim for you | You |
| Accountant | Making schedules reconcile to the accounts | You |
| FlowParse | Turning documents into checkable rows | Free tier, then per page |
The row people misread is the third. A loss adjuster is competent and usually reasonable, and is also appointed by the insurer. Treating them as your advisor leads to under-claiming; treating them as an adversary leads to a defensive file and a slower claim. Professional and complete is the posture that works.
What the adjuster is looking at when your file lands is set out in what loss adjusters look for in your records.
The shoebox problem
Almost every self-assembled claim runs into the same wall: a physical box of paper, a phone full of photographs, and a folder of PDFs downloaded from six different portals.
What makes it hard is not any single document. It is that they are in different formats, describe the same supplier three different ways, and have no common order — so building one table means opening each one and typing what it says, several hundred times, while tired.
That is the part worth removing, and it is genuinely the whole of the benefit. Read into rows, the box becomes a table you can sort, filter, total and search. Suppliers group. Duplicates surface. Missing weeks become obvious because the gap is visible in a date column instead of hiding in a pile.
One practical note about photographs. A crumpled receipt photographed at an angle in poor light is still readable, but with lower confidence — and confidence is shown rather than hidden, so the handful of rows worth a second look are marked. Photograph flat, in daylight, and even that small effort disappears.
Net, VAT and gross — keep all three
Whether a claim is settled net or gross of VAT depends on your VAT position and the policy wording, and it is genuinely a question for your accountant and broker rather than something to decide while typing.
The practical consequence is simple: keep all three figures on every row from the start. A schedule holding only gross cannot answer the net question without every document being reopened, and that is a full rebuild for something that was free to record in the first place.
The same logic applies to dates. Document date, delivery date and payment date answer different questions, and the one a claim turns on is usually delivery — the one least often recorded. Capturing all three costs nothing at the point of extraction and is expensive to derive later.
Where a document shows several VAT rates, they come back separately. A single aggregated tax figure is not enough to work with, and reverse charge or exempt lines are reported as they appear on the document rather than being normalised into something more convenient.
Handing over to a professional
Many policyholders start alone and bring someone in once the scale becomes clear. That is a sensible path, and what you have prepared decides what it costs.
A loss assessor or accountant handed a box of documents will start by doing exactly what you would have done — reading them into a table — and will charge for the hours. Handed a table with source references, they start on the part you are paying them for: judgement.
So the work described here is not wasted if you later get help. It is the input either way. The only question is whether it is done at your cost in your time or at their rate on your invoice.
Signals that it is time to bring someone in: the figure is large relative to the business, the policy wording is ambiguous, several entities or sites are involved, or the first response you receive is a detailed challenge rather than a request for documents.
What to hand over when you do. The schedule, the documents in schedule order, the covering note, and — most usefully — the list of what you already know is missing or estimated. That last item saves a professional a week of rediscovering your own weak points, and it is the thing most people are reluctant to volunteer precisely because it looks like an admission. It is not; it is a head start.
Keep the master file yours rather than handing over the only copy. Whoever helps will work faster from a table they can filter, and you will still need to answer questions about your own business after they have finished with it.
Six mistakes policyholders make
Waiting until asked. The perishable evidence is disappearing while you wait, and it is the only category that cannot be recovered at any price.
Submitting everything. Volume produces sampling, and sampling produces discoveries. A tight, referenced schedule beats a crate of paper.
Skipping the small costs. Increased costs of working are individually trivial and collectively material, and they are what gets dropped when the process is exhausting.
Hiding the gap. A disclosed gap is a fact about the world; a discovered one becomes a fact about your file.
Letting the total move. A figure that changes between versions resets confidence in everything around it. If it must change, change it openly and reissue the whole schedule.
Four people, four spreadsheets. Reconciling them takes longer than gathering the documents did, and the reconciliation becomes its own source of error.
Starting today
If the loss is recent, the order is not negotiable: photographs before clean-up, then thermal receipts, then exports from anything that might be replaced. Those three cannot be done later.
If the loss is not recent and you are already several weeks in, start with the bank statements. They are complete, independent and still available, and they form the spine that everything else attaches to — including the list of movements with no document behind them, which is how you find out what is missing.
Either way, the first pass through the documents is worth doing in one sitting rather than as questions arrive. Not because it is more efficient — though it is — but because seeing the whole shape at once is what tells you which parts of the claim are strong and which need work while there is still time to do it.
The full eight-step sequence, including the baseline work that most claims are actually decided on, is in how to evidence a business interruption claim.