At month-end, every draft invoice for an insurer client is checked against that insurer's billing guidelines: approved rates, unpaid task types, daily hour caps and block-billed or vague narratives. Each lawyer gets a short list of flagged entries to fix before the invoice goes out.
Typical results
≈8 hrs/mo
Billing partner review time saved
100%
Time entries checked before the invoice goes out
Day 1
Flag lists sent on the first business day
Figures reflect typical results for this kind of automation.
The challenge
Insurers that send defense work set billing guidelines, and their bill reviewers cut any entry that breaks them: a rate above the approved one, internal conferences, more than ten hours in a day or several tasks lumped into one narrative. Reading every pre-bill takes the billing partner about 12 hours a month, and about 8% of fees are still cut.
Before and after
Share of billed fees cut by insurer bill reviewers
−63%Billing partner hours a month on pre-bill review
−67%The automation
At month-end, Make pulls the draft invoice for each insurer client from the billing system. Every time entry is checked against that insurer's rules in a Google Sheets table: approved rates, task types it will not pay for, daily hour caps and staffing limits. OpenAI flags narratives that are vague or combine several tasks. Each lawyer gets one list of flagged entries with the broken rule and a suggested fix, and the invoice is released once the list is clear.
How it works
- 01
Draft invoices pulled
On the first business day of the month, the draft invoice and time entries for each insurer client are pulled from the billing system.
- 02
Insurer rules loaded
Each insurer's approved rates, unpaid task types, daily hour caps and staffing limits are read from a Google Sheets table.
- 03
Entries checked
Rates, task types and hours are checked by rule, and AI flags narratives that are vague or lump several tasks into one entry.
- 04
Flag list per lawyer
Each responsible lawyer gets one list of flagged entries with the broken rule and a suggested fix, such as splitting an entry or writing it down.
- 05
Invoice released
Edits in the billing system clear each flag, write-downs are logged by insurer, and the invoice is released once its list is clear.
Starts from
- Month-end draft invoices
- Insurer billing guidelines
Delivers
- Flag list per lawyer
- Invoice released when clear
- Write-down log by insurer
The impact
Fee reductions by insurer bill reviewers typically fall from about 8% of billed fees to 3%, because entries that break a guideline are fixed or written down before the invoice leaves. The billing partner's review drops from about 12 hours a month to 4, spent only on flagged lines.
Questions about this automation
How are each insurer's guidelines turned into rules?
The firm enters each insurer's guidelines once in a Google Sheets table: approved rates per timekeeper level, task types the insurer will not pay for, daily hour caps and how many lawyers may bill for one hearing. When an insurer issues new guidelines, the table is updated and the next month's check uses it.
What does the AI check that the rules cannot?
Narratives. Bill reviewers reject entries such as 'work on file', or a single entry that combines research, a client call and drafting. OpenAI reads each narrative, flags it as vague or block-billed and suggests a rewrite or a split. The lawyer decides whether to accept the suggestion, edit the entry or leave it as it is.
Does the automation change or send invoices on its own?
No. It reads draft invoices and reports what breaks a guideline, and every edit or write-down is made by a lawyer in the billing system. The invoice is released for sending only once its flag list is clear. Simple automations like this are usually live within days, often starting with the two or three insurers that cut the most.