Most legal departments still review outside-counsel invoices in email. A PDF arrives. Someone forwards it for comment. A reply says it looks fine. The invoice moves to finance. The process works well enough that nobody questions it, which is exactly why it keeps costing money.
The status quo is a conversation, not a process
Email review has no structure. Each invoice becomes a one-off conversation with its own participants, its own thread, and its own ending. The knowledge produced by that review, the rates checked, the line items questioned, the reasoning behind the final decision, lives in the thread and dies with it.
That matters because invoice review is not really a reading task. It is a comparison task. You are comparing what a firm billed against what a firm agreed to bill, and against what the matter was budgeted to cost. Email gives you the document but none of the comparison.
Three places the money leaks
The first leak is rate deviation. Your organization negotiated rates. The invoice may not follow them. Catching that in email depends on a reviewer remembering the agreed rate for that timekeeper on that matter, often months after the agreement was signed. Memory is not a control.
The second leak is the missing budget view. Each invoice can look reasonable on its own while a matter quietly runs past its budget. Reviewing invoices one at a time hides the cumulative picture. By the time someone notices, the spend has already happened.
The third leak is approval ambiguity. When approval is a reply, it is unclear who owns the decision. Two people may reply. Neither may be the approver. Nobody can prove later who signed off, on what version, or on what basis.
What changes when the work has a record
Structured review attaches every invoice to its matter. Rates are checked automatically against the agreements already on file. Deviations surface as flags rather than as something a reviewer has to notice. The invoice routes to a named approver, and the decision is recorded against the invoice itself.
The reviewer still decides. Nothing about this removes judgment from the person doing the work. What changes is where their attention goes. Instead of gathering context, they spend their time on the questions that need a human answer: was this work necessary, was it staffed correctly, is the result worth the cost.
The second change is cumulative. Once reviews are structured, the review itself becomes data. You can see which firms deviate most often, which matters run over, and how long approvals take. None of that is available from a mailbox.
Audit your current process
You do not need a project to find out whether this applies to your team. Five questions will tell you.
- Can you list every invoice currently awaiting approval, and who owns each one?
- Are agreed rates checked automatically, or from memory?
- Can you see budget against actual for a matter before the next invoice arrives?
- Could you produce the full approval trail for last year's largest invoice today?
- Do you know your average time from invoice receipt to approval?
If the answer to any of these is no, the gap is structural. It is not a sign that your team is careless. It is a sign that the tool holding the work was never designed to answer those questions.
Start with the invoices you already have
The practical first step is small. Take one month of invoices from one firm. Put them on a structured record with the rate agreement attached. Route them to a named approver. Then look at what the review surfaces that email did not.
Most teams find something in the first month. Not because anyone was acting in bad faith, but because a process built on forwarding and replying cannot check its own work. A record can.
Why careful reading does not fix it
The common response to invoice leakage is to review harder. Assign a more senior reviewer. Add a second pair of eyes. Ask people to slow down. This rarely works, and it is worth being precise about why.
Reviewing an invoice properly means holding several things in mind at once: the rate agreement, the matter budget, the staffing plan, and what was actually delivered. A reviewer working from a PDF has one of those four in front of them. The other three are somewhere else, if they are written down at all. Asking for more care is asking someone to compensate manually for missing context.
It also does not scale. A team can read carefully through twenty invoices a month. At two hundred, careful reading becomes skimming, and skimming finds only the obvious problems. The errors that cost the most are rarely obvious. They are small deviations repeated across many line items.
The reporting you cannot produce
There is a second cost that does not appear on any invoice. Because email review produces no data, legal cannot answer routine questions from finance and leadership.
- Which firms bill above agreed rates most often, and by how much.
- How much of last quarter's spend went to which practice area.
- Average time from invoice receipt to approval, by approver.
- Which matters are trending over budget while still open.