My Top 10 Signs Your Practice Is Ready for an AI-Native Rebuild

A rebuild is not a small ask. It touches pricing, workflow, staffing and how a firm talks to its own clients about what it delivers, and I do not recommend it lightly. Over enough of these conversations, ten signs have emerged that reliably distinguish a practice that is genuinely ready from one that would just be adding AI on top of the same unstructured process, which never works. Here is the diagnostic, in the order I actually check it.

1. You can already name your process, stage by stage

If a partner cannot walk me through their own matter lifecycle without a lot of "it depends" and "usually," the practice is not ready for a rebuild, it is ready for the documentation step that comes before one. AI amplifies whatever structure already exists. A practice that cannot name its own process yet has nothing for the rebuild to amplify.

2. The same work repeats often enough to be worth systematising

A practice handling one bespoke, entirely novel matter every few months has less to gain from an AI-native rebuild than one running dozens of structurally similar matters a year. Repetition is what makes systematisation pay off. Volume is the signal I look for before anything else.

3. Someone in the firm is already quietly relying on AI for real work

If a paralegal or associate is already using a chatbot informally to draft first passes or summarise documents, and getting real value from it, that is a strong signal the underlying work is a good fit for structured AI, not just an ad hoc habit. The informal use is evidence, and I take it seriously as evidence.

4. Billable hour pricing is creating visible friction with clients

When clients are pushing back on invoices, asking for estimates the firm cannot confidently give, or explicitly asking about flat fees, that friction is a sign the underlying economics are ready to change, and AI-native delivery is what makes a genuinely different pricing model possible rather than aspirational.

5. There is one person, not a committee, who can say yes

I will not run a rebuild for a room that cannot commit. A managing partner or sole practitioner who can make the call directly is a green light. A rebuild proposal that has to survive a partnership vote, a technology committee, and a risk committee sequentially rarely survives long enough to actually ship.

6. The firm already trusts a checklist or SOP culture

A practice that already writes things down, has SOPs, uses templates consistently, is much closer to ready than one where every matter is run from an individual lawyer's memory. The rebuild does not have to fight an anti-documentation culture on top of everything else.

7. There is a named person willing to own the system going forward

A rebuild that nobody owns after launch decays within a year, because the process underlying it keeps evolving and nobody updates the system to match. I look for a specific person, not a role description, who is willing to be that owner before I start.

8. The practice has enough matters open at once that a person cannot hold them all in their head

This is a very concrete signal. If a lawyer can genuinely track every open deadline across every matter from memory, the pain a rebuild solves is not yet large enough to justify it. Once that stops being true, and it usually stops being true faster than people expect, the case for a system becomes obvious rather than theoretical.

9. Intake or document assembly is visibly the bottleneck

When I ask where time actually goes and the honest answer is repetitive intake, repetitive drafting, or repetitive document review rather than genuine legal judgment, that is exactly the work AI-native systems are built to carry, and it is a strong, specific signal the rebuild will free up real capacity rather than just adding a new tool to check.

10. The firm is willing to change the process, not just add a tool on top of it

This is the one I weigh most heavily, and it is where most practices that are not ready reveal themselves. A firm that wants AI bolted onto its existing, unexamined workflow, with nothing about the underlying process changing, will not get an AI-native outcome no matter which tool it buys. A firm willing to actually examine and rebuild the process itself, with AI running inside the new structure rather than draped over the old one, is the one a rebuild will actually work for.

What I do when a practice hits only a few of these

Not every practice needs to hit all ten before a rebuild makes sense, but a practice hitting only two or three usually needs a smaller first step, documenting one workflow, systematising one recurring matter type, before a full rebuild is the right recommendation. I would rather tell a firm honestly that they are not ready yet than sell them a rebuild their process cannot actually support. The firms that wait until they genuinely hit most of this list end up with a much smoother transition than the ones who rush it because the idea sounded good in a meeting.