Case Study: A Criminal Defense Firm's Intake, Automated
A multi-office criminal defense firm went from no AI to sign-and-pay retainers and photo-priced traffic tickets in two weeks. What we built and the numbers.
Founder, Vertana · Non-attorney founder and Firm Manager, Conduit Law
Reviewed by Jonathan Mahler
One workflow
Request demo
Bring one recurring bottleneck. If it is a good first firm-brain build, the next step is a live demo, not a pitch deck.
In late October a multi-office Texas criminal defense firm hired FirmOps. They handle criminal cases and a steady volume of traffic tickets. They had no AI subscription of any kind. Within two weeks of the kickoff call, clients were signing and paying for ticket cases from a phone photo with no one at the firm touching the file. This is what we built, what it replaced, and what the numbers say so far.
We aren't naming the firm here. The numbers are real and come from the firm's own portal records.
Where the firm started
Like most firms, the work the firm runs on lived in people's heads and in hand-offs. That is the problem a firm brain is meant to solve. (For the broader map, see AI for criminal defense lawyers.)
- Retainers were built during the call. The attorney quoted a fee by phone, then handed the caller to a paralegal. The paralegal typed the agreement and the payment plan while the client waited. Staff called this the slow part.
- Signing started a second job. After a client signed, staff worked through about 13 steps by hand. They confirmed the payment, marked the lead Hired in Clio Grow, converted it to a Clio Manage matter, set the practice area and attorney, typed in the contact details, looked up the case with the court, filed the letter of representation, uploaded the paperwork, logged the emails and assigned the tasks.
- Tickets ran through a web form that turned work away. The form had one price. It declined commercial drivers, tickets 15 or more over the limit, accidents and school zones outright. It promised the client "a contract within ten minutes". The automation behind that promise was never finished.
- No AI. The firm had no Claude or ChatGPT account. Staff had never had AI work inside Clio or the firm's email.
What we built
1. A firm brain in Claude
Staff got Claude, connected to the systems the firm already pays for: Clio Manage, Clio Grow, Dropbox, CallRail, Gmail drafts and Google Drive. That is 46 tools behind one sign-in. Staff can ask about a matter, a lead, a call or an email thread, and the answer names the record it came from. Every write runs through a preview a person confirms before anything changes. That covers notes, tasks, calendar entries and bills. Nothing goes to a client on its own.
Within days, a member of the firm's staff wrote to us:
"I am using the desktop app and have already had it automate a biweekly email task that took me hours every two weeks. My mind is blown!"
We didn't build that one. They did, on the firm brain, on their own. That is the point of putting the firm's systems inside the AI staff already use, instead of buying another tool with its own login.
2. A sign-and-pay retainer link, sent during the call
Staff now enter the case type and the fee while the client is still on the phone. The client gets a text with the agreement already filled in. They sign and pay through Clio Payments in the same flow. When they finish, the Clio contact, the matter and the signed PDF are created automatically. The $1 e-signature cost is posted to the matter as an expense. The 13-step checklist after signing is now the system's job, not a paralegal's.
3. Traffic tickets from a photo
The client takes a photo of the ticket. The system reads the court, the citation numbers and each violation. It prices them from the firm's own fee schedule, which we turned into 40 pricing rules, 32 of them fully automatic. Then it sends the client straight to sign and pay.
Not every ticket should be priced by a machine, and the firm decides which ones. We mapped 127 courts. The firm takes 95 automatically, declines 16, and sends 16 to a person. A ticket that needs judgment goes to staff with the photo, what was read and the reason. Staff set the price and the client gets the same link to continue. Commercial drivers and accident tickets now get a price or a call from the firm. The old form just said no.
4. The back office after the signature
Once a ticket client pays, the system drafts the letter of representation for that court. A paralegal sends it from the portal by email, fax or mail, depending on what the court takes. We recorded which method each court uses when we built the court list.
5. A staff board and one lead list
Staff get one board showing every retainer and ticket link: sent, opened, signed, paid, or stuck. Calls from CallRail and leads from Clio Grow go onto one running list. A new call is matched to the lead it belongs to by phone number, or starts a new lead.
The numbers so far
Tickets cover the first 15 days after launch. Retainers cover the first 10 days of the link.
| Traffic tickets | Criminal retainers | |
|---|---|---|
| Links started | 155 | 18 |
| Priced | 72 (46 with no staff input) | Set by staff |
| Signed | 44 | 7 |
| Paid | 36 | 5 |
| Median time, start to signed | 15 min (about 5 when priced automatically) | About 2 hours |
Of the 44 signed ticket clients, 32 signed within half an hour of starting. 23 reached a signed agreement without anyone at the firm touching the file.
Time saved (our estimate). Under the old process, each ticket meant reading the form, preparing the contract, taking the payment and working the Clio checklist. We put that at about half an hour of staff time per ticket, which is conservative. At 43 signed tickets in about 15 days, that is roughly 20 hours of staff time. That's about a week and a half of one person's month on tickets alone, before counting retainers or the hours staff are saving themselves in Claude.
What we would tell another firm
- Start where the money changes hands. The sign-and-pay step had the most hand-offs and the most waiting. It was the first thing we fixed, and it is where clients notice the difference.
- Write the firm's judgment down, then automate it. The ticket flow works because the firm told us which courts and violations to take, decline or send to a person. The machine only applies those rules.
- Keep the people who approve. Every write in the firm brain is previewed and confirmed. That is what made it safe to connect Clio and the firm's email in week one. More on that in read-first AI.
- Put the AI where staff already are. The staff member who automated their own two-week email task didn't open a new app. She used Claude, with the firm's systems already connected.
What's next
Next up: texting clients from the firm brain through CallRail, filing signed agreements into the firm's Dropbox, and asking happy clients for reviews at the right moment. We will update this page with the numbers when each one is live.
FAQ
Which practice management software does the firm use?
Clio Manage for matters and billing, and Clio Grow for leads. Everything here writes to Clio through Clio's API. The firm didn't change systems.
Does the AI send anything to clients on its own?
The firm brain doesn't. Every write is a preview a person confirms. The sign-and-pay links go out when staff send them, or when a client starts one themselves from the firm's website.
How long did it take?
The firm brain was live four days after the kickoff call. Ticket sign-and-pay was live eight days after it, and the first paying ticket client came two days later.
Can a firm without traffic tickets use this?
Yes. The ticket flow is one use of the same sign-and-pay link the firm uses for criminal retainers. We run the same pieces at Conduit Law, a personal injury firm. See how we work with firms on law firm operations.
Keep moving up the operating system
Get the Managed Firm Brain Notes
Follow the practical path: live workflow builds, CRM-aware automation when the data is ready, and supervised work after approval.
Demo Context
What to watch for in the Managed Firm Brain walkthrough
Live-Lab Proof
How Conduit informs the first external deployments
Fit Criteria
API-ready systems, workflow scope, and approval gates
About Jonathan Mahler
Jonathan Mahler is the founder of Vertana (FirmOps is a Vertana brand) and the non-attorney founder and Firm Manager of Conduit Law. He runs the systems of a live PI firm every day, then turns reusable patterns into practical Managed Firm Brain workflows: approved context, supervised drafts, approval gates, and a path into deeper automation when the first workflow proves value. Meet Jon and Elliot.
Related Articles
What is a firm brain?
A firm brain is a purpose-built operating layer for one law firm, implemented in partnership, then run for that firm. Not a chatbot. Not a boxed product.
AI Approval Gates for Law Firms: What Stays Human
Firms buy AI and skip the approval layer. A PI firm operator maps the two gates that run in production, who holds each one, and the minimum version to retrofit.
Read-First AI for Law Firms (Before You Automate a Send)
Vendors sell the send. Operators start with visibility. The read-first audit on Clio, email, and intake, what disqualifies a firm, and the gates that stay human.
Request demo
Bring one recurring bottleneck and see whether it is a good first firm-brain build. If the first workflow is concrete, the next step is a live build session, not a broad advisory project.
Live demo | Human approval before external action