AI pre-bill review: catch the write-down before the client does

Every partner who has sent bills to an insurance carrier or a corporate client with outside-counsel guidelines knows the ritual. The bill goes out. Weeks later it comes back reduced, with a list of reasons: narratives too vague to evaluate, block-billed entries lumping four activities into one line, task codes that don't match the work described, hours the guidelines say should have been staffed differently. The firm eats the difference, someone spends an afternoon arguing about a few of the lines, and next month it happens again.

The frustrating part is that almost none of this is a dispute about the work. The work was done. The bill just failed to say so in the way the reviewer — human or algorithmic — needed to see it. That makes it an engineering problem, not a client-relations problem: the defects that get bills discounted are identifiable before the bill leaves the firm, which means they're fixable while they're still yours to fix.

Why bills actually get discounted

The recurring offenders are boringly consistent:

• Vague narratives — "attention to file," "work on motion." A narrative that gives the reviewer nothing to evaluate gets cut, whatever the work behind it was.
• Block billing — one line, several activities, one time figure. Because no single activity can be assessed, the whole line is vulnerable.
• Coding problems — the right work under the wrong task or activity code reads as noncompliance even when the underlying time is legitimate.
• Rate inconsistencies and hours patterns that look excessive against the guidelines, which draw the same treatment.

Here's the framing that changes behavior: every one of these defects is visible in the draft bill. The information needed to catch them exists in-house, weeks before the client sees anything. A firm that reviews for these defects systematically — every bill, every line — converts write-downs into edits. The only question is whether anyone has time to do that review by hand. Mostly, nobody does. So the reviewer of first impression ends up being the client's auditor, which is the most expensive possible place to discover a vague narrative.

And the cost isn't only the discounted lines. Every reduction spends a little of the client relationship — the conversation about the bill displaces the conversation about the case — and disputing the cuts costs partner time the firm will never bill to anyone. The cheapest version of every one of these defects is the one fixed in the draft.

Two reviews before any client sees the bill

Guideline validation: per-line violations, before sending

EdgeLex makes the review systematic. Every bill validates against billing guidelines and outside-counsel rules before it goes out — producing per-line violations and a compliance score. The block-billing and coding problems carriers discount for get caught in-house first, flagged on the specific lines that carry them, while the bill is still a draft.

Think of it as the same review the client's auditor will run, run early, by your side. The compliance score tells you at a glance whether a bill is ready; the per-line flags tell the billing attorney exactly what to fix. No more discovering in the remittance advice which lines failed — you knew before you sent it, and you fixed the ones worth fixing.

The AI pre-bill pass: a second reader for every draft

Guideline validation catches rule violations. The AI pre-bill review reads the bill the way a demanding reviewing partner would. It makes a pass over the draft and returns a quality score with flagged findings: vague narratives — with suggested rewrites, not just complaints — rate inconsistencies, and excessive-hours patterns. Specific fixes, attached to specific lines, before a client ever sees the bill.

The defects that get bills discounted are visible in the draft. The only question is who finds them first — your firm, or your client's auditor.

The suggested rewrites matter more than the flags. Telling a timekeeper their narrative is vague is a critique; handing them a rewrite that says what the entry actually accomplished is an edit they can accept in seconds. It's the difference between a reviewer who marks up your draft and one who fixes it — and it's why a pre-bill pass that would take a billing partner an evening happens in the ordinary flow of monthly billing.

The plumbing underneath: codes, formats, and honest invoices

None of this works without structured time capture, and EdgeLex treats that as foundation rather than afterthought. Time entries carry UTBMS activity and task codes, with timers that follow the work, rate schedules per engagement, and expense tracking. Bills generate from unbilled time, and the line items can be worked — edited, discounted, written off, re-coded — with running totals that always reconcile to the invoice.

Capture itself gets AI help too, which matters because the worst narratives are usually the reconstructed ones — the entries written Friday afternoon for work done Tuesday. EdgeLex can capture time and expenses from activity and auto-generate billing narratives, so the entry gets written closer to the work, from the work. And on mobile, capture follows the lawyer: structured time entry, a matter-guarded billable timer, receipt-OCR expense capture, and the option to simply describe the work to Lex and approve the entry it drafts. The less reconstruction in the timekeeping, the less rewriting in the pre-bill review.

Output meets the client where they are. LEDES export feeds e-billing systems; PDF and themed invoice templates — with a live preview before anything is sent — cover the clients who are people rather than audit platforms. Same underlying bill, both audiences, and you see exactly what the client will see before it goes.

Approval, versioning, and finalize-to-lock

A bill is a representation to a client, and EdgeLex treats it with the corresponding discipline. Drafts move through an approval workflow — draft, review, approved, finalized — with revision reasons recorded and bill versioning throughout. Finalizing locks the bill: the invoice a client receives can never silently change afterward.

That lock earns its keep the day a client questions an invoice from eight months ago. The bill they received is the bill on file — frozen, versioned, with the revision history that led to it. Delivery goes by email, mail, or portal with view tracking, payments get recorded, payment plans run with dunning, and collection risk surfaces before it becomes a write-off.

Trust accounting, done properly

Billing in a law firm ends at trust, and this is the part where "close enough" is not a category the bar recognizes. EdgeLex carries client trust ledgers with deposits, disbursements, and reconciliation — the accounting a bar audit expects, not a spreadsheet approximation. Retainers are managed on the same rail, including evergreen arrangements that auto-replenish.

The point of having trust accounting inside the same system as time, billing, and collections is that the money's whole story stays connected: the retainer collected at engagement, the invoice it was applied against, the reconciliation that proves the ledger. One system of record, one audit trail, and firm-controlled data throughout. Generic billing tools don't understand trust obligations, and the workaround — a spreadsheet beside the billing system — is precisely the arrangement that makes a bar audit longer than it needs to be.

What changes when the review is systematic

Run the sequence end to end and the shape of billing changes. Time gets captured structured, with the codes the guidelines demand. The draft bill gets two readers before any client does — a rules engine producing per-line violations and a compliance score, and an AI pass producing a quality score, flagged findings, and suggested rewrites. The billing attorney works a short list of specific fixes instead of proofreading pages of entries. The bill goes out clean, in the format the client's systems expect, and it locks on finalize.

Nothing in that sequence required anyone to work more hours. It required the review that was always worth doing to stop depending on somebody finding an evening for it. That's what pre-bill review is for: the firm keeps revenue it already earned, and keeps the version of the client conversation that isn't about the bill.