The sixty-second answer
Read the letter and work out precisely which years and which items are in scope - it is usually narrower than it feels. Note the deadline. Gather the supporting records for those specific items, send readable copies rather than originals, and say plainly if something is missing. The hard part is assembly, not argument.
The letter is narrower than it feels
Correspondence from a tax authority produces a predictable reaction, and that reaction is not useful for the first ten minutes. The productive first step is mechanical: identify the tax years named, and identify the specific items being queried.
A great deal of correspondence is not a full audit. It is a targeted query about particular claims - a vehicle expense, a category that looks unusual against the rest of the return, a credit claimed. These are resolved by sending the records that support the claim.
An audit proper is broader. An authorised person may inspect, audit or examine books and records, and any document that relates or may relate to what should be in those records [2]. But even then the process starts with specific requests, and each is answered the same way.
Why gathering is the real work
In almost every case the owner already knows the answer. They know the vehicle was for work, that the tools were for a job, that the meal was with a client. What takes the time is producing the evidence.
That is a filing problem rather than a tax problem, and it is decided long before any letter arrives - by whether receipts were captured at the time, whether they are searchable by supplier and date, and whether the images are readable years later.
Records generally have to be kept until six years after the end of the last taxation year they relate to [1]. A request three years back is entirely ordinary, which means the relevant question is whether your 2023 records are findable now, not whether they existed then.
Electronic records are fine, unreadable ones are not
Electronic records are permitted, provided they are retained in an electronically readable format [1]. Scanning and photographing receipts is a legitimate approach.
The failure mode is not format, it is legibility. Thermal paper fades - sometimes within a year - and a photograph taken of an already-faint slip does not improve with age. A receipt that cannot be read supports nothing, whether it sits in a folder or a database.
Capturing at the point of receipt, while the print is still dark, is the difference between a record and a picture of a blank rectangle.
Sorting a pile against a specific request
The practical bottleneck is matching a request to the documents that answer it. A request names categories and periods; a shoebox is organised by nothing at all. Bridging that gap by hand is where the days go.
If the records are already categorised and searchable, the bridge is a query - this supplier, these months, this category. That is the specific job Penny does inside MapleExpense: classifying documents as they come in and pulling together the ones that relate to a particular request, so the response is assembled from an index rather than from memory.
It does not decide the tax question. It removes the week of sorting that otherwise sits in front of the tax question.
Missing records, handled honestly
Some documents will be gone. Say so, and provide the best alternative evidence available - a supplier copy, a contract, a bank record combined with correspondence establishing what the payment was for.
Alternative evidence is a weaker position and sometimes it will not carry the claim. That is a worse outcome than having the receipt. It is a far better outcome than creating a document after the fact, which moves the matter out of the realm of documentation disputes entirely.
When to bring in a professional
For a narrow request about a few claims, most owners can respond themselves. The calculus changes when the request spans several years, when the adjustment at stake is large, or when you disagree with the position being taken rather than simply needing to evidence it.
Those are matters of judgement and negotiation, not filing, and they are worth paying for. Sending the receipts is not.