The consult ended well. The agreement went out three days later.

The consultation went well. They said they wanted to move forward. Then the day filled up, the notes sat there, and the representation agreement went out three days later. To someone who had spent those three days talking to another firm. That gap is where firms lose people who had already decided to hire them. This is about one firm's fix for it. Not the software part. The part before the software, which is where almost every attempt at this falls apart. ## What we were actually trying to fix The firm's own description of the problem, roughly: lawyers check their meeting notes, draft the agreement, and send it when they get a minute. Sometimes that minute is hours later. Sometimes it is days. Staff re-read the same notes twice because nobody is sure what was agreed. And some people who said yes in the room never sign anything, because by the time the paperwork arrives the moment has passed. The goal was narrow: when someone says yes in a consultation, the agreement reaches them while they still feel like they said yes. ## Write the journey before you build it The first version of this work happens on a whiteboard, not in a tool. Two passes. **The first pass is the journey, end to end.** Someone books a consultation. The consultation happens. There is a decision: do they want to proceed? If yes, the agreement goes out, gets signed, and their record changes from lead to client. If no, they go into follow-up instead of disappearing. That is it. Ten boxes, and everyone in the firm can read it and say whether it is right. If the partners cannot agree on that drawing, nothing built on top of it will work either. **The second pass breaks each box into the actual actions.** Take "book a consultation." Underneath it: they land on the site or click an ad, they fill out a form or pick a time on the calendar, their details get saved, their record gets marked as a new enquiry, they get a confirmation, and if there is a meeting on the books they get reminded before it. Those small steps are the build. The first drawing is for the humans. The second is what gets built. ## The part that trips people up You cannot instruct software the way you instruct a paralegal. "Book a consult, then run the consult" is a sentence a person understands and a machine cannot act on. Everything has to be *when this, then that*. So the second drawing reads more like: when a form comes in, create the contact. When a consultation is marked complete, look at what was agreed. When the answer is clear enough that they want to proceed, send the agreement for signature. That third one is the one to argue about, and it deserves the argument. Deciding "clear enough" is a judgement call that a firm is handing to a machine. Get that threshold wrong in one direction and agreements go out to people who were still thinking. Get it wrong in the other and you are back to sending them by hand. This is not a detail to settle during the build. Settle it with the person whose name is on the letterhead, before anything is switched on. ## The honest limit Mapping does not make a bad process good. If the agreement takes three days because nobody has decided who is allowed to approve one, drawing it will not fix that. It will just show you, on paper, exactly where the three days go. That is still worth doing. But the answer will be a decision somebody has to make, not a piece of software. And there is a second thing the drawing will not cover: what happens when the agreement is sent and never signed, or when half the details are missing. Every real version of this ends up being mostly exception handling. Plan for that being the bulk of the work, not an afterthought. ## What actually changed The order matters more than the tools. Understand the journey, then break it into steps, then build. Doing it in the other order is how firms end up with automation nobody trusts and everybody quietly works around. And the thing worth protecting is not the time saved. It is the fact that the person who said yes gets treated like they said yes, on the same day.