law firm client intake process: the steps, and where each one breaks
the short answer
a law firm's client intake process should run in six steps, each with an owner and a clock:
- capture. every call, form, chat and email lands in one list, with the time it arrived.
- screen. practice area, jurisdiction, what happened and when. check the deadline first, because a statute of limitations can close a case before the firm calls back.
- conflict check. the caller's name and the other side's name, against the firm's records, before any consultation is booked.
- decide. a lawyer reviews the facts and says yes, no, or "we need X first". software can grade the facts. it should not make this call.
- engage or decline. an engagement letter for a yes. a written decline for a no, which also protects the firm if the person later says they thought they were a client.
- open the matter. create the file, start collecting documents and medical records, and hand it to the case team with everything gathered so far.
playground giants, jimmy hackett's consultancy in houston, builds this flow on a firm's own accounts. here it is as a chart, with where each step usually breaks.
the flow, step by step
| step | what goes in | what comes out | who owns it | where it breaks | |---|---|---|---|---| | 1. capture | a call, form, chat or email | one row with the source and time | intake staff, or the system | after-hours calls ring out; forms land in an inbox nobody watches on weekends | | 2. screen | the row | practice area, jurisdiction, dates, injury, deadline | intake staff | screening from memory, so two people grade the same caller differently | | 3. conflict check | both parties' names | clear, or a flag | intake or records | done after the consultation, when the firm has already heard the facts | | 4. decide | the screened facts | yes, no, or "need more" | a lawyer | nobody writes down why, so the next "no" is argued again from scratch | | 5. engage or decline | the decision | a signed engagement letter, or a written decline | a lawyer or intake lead | the letter goes out days later and the person has already signed elsewhere | | 6. open the matter | the signed letter | a file, a records request, a named case team | case manager | facts from intake get typed in again, and some are lost |
where it actually breaks
the steps above are the textbook. here is what i found when i audited one personal injury firm before building anything for it, because the textbook didn't predict most of it.
the problem wasn't the decision step. it was step 1. the homepage had no inquiry form at all, and each page had a single phone link near the very bottom. paid traffic would have arrived, found nothing to press, and left. a practice page also described a serious-injury threshold in language that was no longer accurate, which a prospective client could have relied on. none of that is a flow-chart problem, and all of it had to be fixed before the flow was worth automating.
once it was fixed, the measured version looks like this: a graded lead reaches the firm in under 10 seconds from a form submit, every call lead arrives as a card that calls back from the same local number the person dialed, and the firm picked up its first two paid leads in under 15 minutes. the grading was tested against 66 made-up inquiries written to the firm's own rules, and a daily list shows every inquiry where the software's grade and a person's grade disagree.
what to automate, and what to leave alone
automate the clock, not the judgment. capture, screening questions, the conflict lookup, reminders, document requests and the log are all safe to automate, and they are where most of the time goes. step 4 is not. a model can pull the facts out of an inquiry and ordinary code can check them against the firm's written rules, but a lawyer says yes or no.
if you want the longer version of that line, it's in legal intake automation that grades facts and never merit. if you want someone to test your firm's flow the way a prospect would, after hours and on a weekend, that's the law firm intake audit.