Every month the same stack lands on a partner's desk: a run of draft bills to approve before they go to clients. And every month, approving them turns into detective work.
One entry's narrative just says "attention to file." A matter has time nobody logged until the day before close. A third has months of unbilled work sitting in WIP with no note explaining why it is still there. A fourth reads fine inside the firm but will bounce straight back from the client's billing guidelines. The partner is meant to be making judgment calls about what to bill, what to adjust, and what to write down. Instead they are reconstructing what actually happened, one entry at a time, a few days before the deadline.
If you run billing or legal operations, or you are the partner staring at that stack, the fix that helps is rarely a new billing system. It is turning the month-end scramble into a review a partner can trust in one pass.
The job is to make the bill trustable before a partner signs it
Deciding what to bill, what to write down, and how to describe the work to a client is professional judgment, and it stays with the partner. Some of it is privileged. None of it should be handed to software. The part that quietly eats a partner's week is narrower: getting every time entry, every narrative, and every WIP signal in front of the reviewer clean enough that the judgment is quick.
Done well, a partner opening the pre-bill can see at a glance which entries are missing, late, vague, or duplicated; which narratives will bounce off the client's billing guidelines before they are sent; and which matters have WIP aging quietly toward a write-down. What the review does not do is set rates, decide what is billable, or put a value on the work. It makes sure the partner is judging facts, not hunting for them.
Why the pre-bill turns into detective work
Billing data is scattered by design. The time sits in the practice management or billing system. The context that explains a thin entry sits in the matter file and the fee earner's memory. The client's billing rules sit in an outside counsel guideline document that nobody has open while they draft. Finance has one view of what is ready to bill, the fee earner has another, and the partner has a third.
So the close becomes a relay of chasing. Someone chases the missing time. Someone rewrites the thin narratives. Someone digs up why a matter's WIP is high and whether it can still be recovered. Someone rebuilds the whole picture for the partner by hand, a few days before the deadline. Next month it all happens again, because none of the answers were written down anywhere the next person can find them.
How the month-end pre-bill actually moves today
In most firms, the pre-bill run looks something like this.
- Finance generates the draft bills for the period and sends them, or a link to them, to each billing partner.
- The partner opens their matters and starts reading. Some narratives are clear. Many are not.
- Thin or missing entries go back to the fee earner, who reconstructs what they did from a calendar, an inbox, and memory.
- Someone notices a matter with high WIP and asks why. The answer is usually in one person's head.
- A bill under a strict client's guidelines gets a closer look for block billing, non-billable tasks, or wording the client's reviewer will reject.
- The partner makes write-down and hold decisions, often without the full story, because the deadline is closer than the answers.
- Finance finalizes, sends, and waits to see which bills the client's portal bounces back.
None of this means anyone is careless. It means the work of making a bill reviewable has never been designed as its own step. Each person sees only their part of it, and the partner inherits the gaps.
Where the pre-bill breaks
The breakpoints are usually specific, and they repeat every month.
Narratives are too thin to bill
"Attention to file", "review documents", "call with client", "consider matter": entries like these tell the partner little and tell the client's reviewer even less. The fee earner knew exactly what they did at the time. Two weeks later, at close, that knowledge has to be rebuilt from scratch.
Related time is scattered or block-billed
Work on one workstream shows up as a dozen small entries across three fee earners, or as one long entry that lumps five tasks into a single line. Many corporate clients reject both. Their guidelines want each task described and timed on its own, the way the UTBMS task codes used across legal e-billing are meant to work.
Time is late or missing
Time entered days after the work, or not at all until the close, is where realization leaks first. The entry is vaguer because it is written from memory, and the WIP looks wrong because half of it arrived at the last minute.
WIP is aging with no explanation
A matter carries a large unbilled balance and nobody at the review knows whether it is a milestone not yet reached, a scope conversation nobody has had with the client, or simply time that should have been billed months ago. Without the reason attached to the matter, the partner is guessing, and aged WIP is exactly where write-downs come from.
The client's guidelines are invisible at the point of review
Each major client has its own outside counsel guidelines: time increments, staffing limits, non-billable categories, narrative format, block-billing rules. At the moment the partner is reviewing, those rules are in a PDF on a shared drive, not next to the entry they apply to. So the bill goes out looking fine and comes back reduced.
Nobody owns the rewrite
When a narrative needs fixing, it is often unclear whether the fee earner, the billing coordinator, or the partner's assistant should do it. So it waits, and it lands back on the partner at the worst possible moment.
A quick test before you build anything
Pull one matter heading into billing and ask whether you can see, without calling anyone, which entries need a rewrite, why the WIP is where it is, and what the client's guidelines will reject. If answering that takes three conversations and two exports, the work of making the bill reviewable is not done. It is just hidden, and it will surface on the partner's desk under deadline.
What good looks like before a partner signs
A better pre-bill does not need perfect time entry or a new billing platform. It needs a small number of things to be true before the partner opens the matter.
The narratives that fall short of the client's standard are already flagged, with a cleaner draft attached for someone to approve. The matters with aging or blocked WIP carry the reason they are sitting there. The client's specific billing rules are visible next to the entries they judge. And it is clear, in advance, which matters can move straight through and which need a partner's real attention. Put those together and the review stops being a hunt and becomes a short list of clear exceptions, each one already carrying the context needed to decide it.
The billing review queue, and what it holds
The practical way to make that real is a single view of the matters heading into a close, with the entries that need attention already flagged, the matter context attached, and a draft note the fee earner or partner edits rather than writes cold. Instead of reading every line on every bill, the partner works a short queue of exceptions.
A first version usually holds something like this.
| Matter | Billing signal | Why it needs a look | Suggested action | Owner |
|---|---|---|---|---|
| Corporate advisory | Narrative reads only "review documents" | Client guidelines require the task and its purpose | Fee earner rewrites from attached draft | Associate |
| Litigation | WIP over ninety days, no recent client update | Partner needs the progress story before the bill goes out | Attach context, hold for partner review | Matter partner |
| Transaction | Twelve small entries on one workstream | Likely to read as block billing and be reduced | Group and check against guidelines | Billing coordinator |
| Employment | Partner time logged against an admin task | Client does not pay partner rate for this category | Reassign or write down before sending | Partner |
The columns matter more than the rows. Every exception arrives with why it was flagged and who is meant to act, so the review becomes a set of decisions rather than a research project.
Fix the narrative standard before you automate anything
The temptation is to point AI at the time entries and let it rewrite them. That gets the order wrong. Before anything rewrites a narrative, the firm has to agree what a good narrative looks like for that client, because "good" is defined by the client's guidelines, not by a general sense of tidiness. A narrative that satisfies one client's reviewer will be rejected by another's.
Most of the standard is concrete. The entry should name the task, the subject, and the purpose, in the client's required format, without lumping unrelated work together. Written down once per client, that standard is what a person or a model can check an entry against.
| What was logged | Why it fails the client's guidelines | What a reviewable version looks like |
|---|---|---|
| "Attention to file" | No task, no subject, no purpose | "Review and annotate revised share purchase agreement, clauses 4 to 7, ahead of completion" |
| "Emails" | No recipient, no subject, unclear if billable | "Correspondence with counterparty's counsel on disclosure schedule timing" |
| "Call; review docs; update tracker; internal discussion" | Block billing across four tasks in one entry | Four separate entries, each with its own task, subject, and time |
| "Research" | No issue named, reads as open-ended | "Research on enforceability of restrictive covenant under governing law, for advice note" |
The exact wording above does not matter. What matters is that once the standard exists in writing, checking a thousand entries against it is exactly the kind of pattern work that can be flagged and drafted quickly, with a fee earner approving the final wording every time.
Reading WIP before it turns into a write-down
Aged WIP is the other half of the pre-bill, and it is where the real money moves. Unbilled time and disbursements sitting on a matter are only worth what the firm eventually collects, and the longer they sit without explanation, the more likely they end in a write-down. Lockup, the combined drag of unbilled WIP and unpaid bills, is a number most managing partners watch closely, and it is built one un-reviewed matter at a time.
The review cannot decide what to write down, that is the partner's call, but it can make sure the partner is deciding with the reason in front of them. A matter with high WIP should carry whether it is waiting on a milestone, whether scope has quietly grown beyond the estimate, whether there is a client conversation nobody has had yet, or whether it is simply time that should have been billed two months ago. Each of those leads to a different decision, and none of them should be reconstructed live in the review.
Decide which matters need a real look
Not every matter deserves the same scrutiny, and treating them as if they do is how the important ones get rushed. A small fixed-fee matter with clean entries does not need what a high-WIP dispute under strict client guidelines needs. The useful move is to decide, up front, which matters can move through with a light check and which get held for a close read, and to make that call visible in the queue rather than leave it to whoever is reviewing on the day.
| Matter type | How it usually bills | What the review should actually check |
|---|---|---|
| Fixed fee, on scope, clean entries | Straight through | Confirm scope not exceeded; light narrative scan |
| Time-based, familiar client, moderate WIP | Light check | Narrative quality and any entries that breach the client's format rules |
| High WIP, aging, or over estimate | Held for partner | Why WIP is high, whether it is recoverable, whether a client conversation is needed first |
| Strict outside counsel guidelines | Held for close read | Block billing, staffing and rate limits, non-billable categories, narrative format |
Making the scrutiny level explicit is what lets a partner spend the review where the money and the risk actually sit, instead of reading every bill at the same depth and running out of time before the hard ones.
Where AI helps, and where the partner still decides
Narrative and WIP cleanup is where AI earns its place, because most of it is pattern work against a known standard. It can spot the thin entries, group time that belongs on one workstream, pull the matter context that explains a line, check an entry against the client's guidelines, and draft a cleaner internal version for someone to approve. Done against a written standard and a review queue, that removes most of the manual reconstruction.
What it does not do is finalize the bill. It does not set a price, it does not override a partner's write-down call, and it never produces client-facing wording that goes out without a lawyer approving it. Anything touching privileged material or legal judgment stays with the people who carry that responsibility. The draft narrative is a first pass a fee earner corrects, and the data behind it has to keep AI drafts and approved bills clearly apart, so nothing a model wrote can be mistaken for a bill someone signed.
The data and systems this touches
A working queue reaches into the systems that already hold the pieces: the time and billing system, practice management, finance exports for WIP and aged balances, the matter files that carry context, and the client billing guidelines themselves, with tasks and email around the edges. You do not connect all of it at once. You connect the minimum that makes one practice group's or one close cycle's queue genuinely useful, then widen from there. The one firm rule that cannot bend is keeping draft support and final billing decisions clearly separated in the data, so an AI-drafted narrative is never one step away from being sent as an approved bill.
A worked example
Say a 30-lawyer firm where month-end pre-bills sit with partners for about two weeks before they go out. Everything here is illustrative, not a real client, but the shape will be familiar.
A billing partner opens forty matters for the period. Reading them the usual way, roughly half have at least one narrative too thin to send, three carry WIP over ninety days with no note explaining it, and two bills for the same corporate client contain block-billed entries that client's portal has rejected before. Finding all of that takes the partner most of two evenings, several messages to fee earners who are themselves reconstructing from memory, and a couple of exports from finance to work out which WIP is which.
Now run the same forty matters through a queue first. The thin narratives are already flagged, each with a suggested rewrite drawn from the matter file for the fee earner to correct. The three aging matters carry their reason: one milestone not yet hit, one scope conversation overdue, one that should simply have been billed in the prior period. The two block-billed bills are marked against that specific client's guidelines, with the entries grouped ready to check. Instead of forty bills to reconstruct, the partner works a list of maybe a dozen real exceptions, each carrying the context to decide it. The judgment calls, the write-downs, the client wording, are all still the partner's. The detective work is what went away.
What tends to go wrong
A few mistakes show up whenever firms try to fix this.
Treating it as a finance cleanup
Fee earners and partners live with the pre-bill too. If the fix only works for the billing team, the narratives never actually improve, because the people writing them were never part of it.
Automating rewrites before the standard exists
If nobody has agreed what a good narrative looks like for a given client, a model will produce fluent entries that still get rejected. The standard has to come first, per client, in writing.
Letting the reasons live in memory
When the reason for a write-down or a hold stays in one person's head, next month the same question gets asked again. The reason has to travel with the matter, or the queue rebuilds itself from zero every close.
Chasing perfect time capture first
Better contemporaneous time entry helps, and it is worth pursuing, but it is a slow cultural change. The review can add value now, working with the entries you actually have, rather than waiting for a habit that takes a year to shift.
A practical first close cycle
You do not need a program to start. You need one close cycle, one practice group, or one demanding client, and a staged first pass.
| Stage | Focus | What should exist by the end |
|---|---|---|
| Week 1 | Trace how one close actually moves | A map of where time, narratives, client rules, WIP, and partner review sit today, and the queue's first columns |
| Weeks 2 to 3 | Connect the minimum data and set the standard | Time, billing, matter, and guideline data feeding one queue; a written narrative standard for the first client |
| Week 4 | Run one close through the queue | The partner reviews a short exception list, with flags, WIP reasons, and draft rewrites already in place |
The first cycle will not produce a finished system, and it does not need to. What it should produce is one close that runs with far fewer blind spots, and enough of a queue that the second close is easier than the first.
How to tell in month one whether it is working
The signal is what the review feels like. If the partner is approving a short list of clear exceptions instead of reconstructing the month, and fewer bills come back reduced from the client's portal, keep going. If the narrative standard or who owns the rewrite are still unsettled, more automation will not save you, and it is worth fixing those before adding anything on top.
The rule for cleaning up pre-bills
Do not ask, "How do we make our lawyers enter better time?" That is a real problem, but it is too slow to help this month's bills. Ask instead, "What would let a partner trust this pre-bill in one pass?" That question points at the narrower work that pays off immediately: flag the entries that fail the client's standard, attach the reason to aging WIP, put the guidelines next to the bill, and hand each fix to the person who should make it. Better time entry follows more easily once the review stops absorbing everyone's month-end.
How Ubisar would implement this workflow
In week one, we take one close cycle, practice group, or client and trace how time, narratives, billing rules, WIP, and partner review actually move today. The first thing you get back is a billing review queue: the narrative flags, the source context, the WIP reason, and the owner and action for each line, built around your matters rather than a generic template.
Over the next two weeks we connect the minimum time, billing, matter, and guideline data to make that queue real, and let AI flag the unclear narratives, group related entries, and draft the internal rewrites, while fee earners and partners still approve every word and every billing call. By week four, one close should run through the queue with far fewer blind spots, and you should be able to point at specific bills that went out cleaner or held back a write-down. This work fits the AI, Data & Tech Implementation Service, and it sits right next to client status reporting for legal matters, since the same WIP and matter progress feed both.
If the month-end review is costing your partners days they would rather spend on clients, send us the one close cycle that hurts most and we will map the first queue with you.