Law firm intake automation: replies drafted for same-day review, with the conflict check run first

Where intake time goes

Every new matter starts as an inquiry, whether a phone call, a web form, an email or a referral. Someone reads it, asks the screening questions, enters the prospect and the other parties in the practice management system, runs the conflict search, gets a lawyer to clear it, books the consult and later drafts the engagement letter. Each step is small. Together they decide whether the prospect hires you or the next firm on the list.

Clio's 2024 Legal Trends Report measured how firms respond. In a secret shop of 500 firms, 33% responded to emails, down from 40% in 2019, and 40% answered the phone, down from 56%. Nearly half, 48%, were essentially unreachable by phone. Those numbers come from one vendor's study, so treat them as a benchmark to test against your own phones and inbox.

Much of the time goes into rekeying and waiting. The same names get typed into the intake form and again into the matter record. The lawyer who must clear a hit is in court. The prospect hears nothing for a day and calls someone else. This guide is general information, not legal advice, and the responsible lawyer must check the rules and orders that govern each matter.

What a good conflict search covers

A conflict search is only as good as the names that go into it. For each inquiry the search should cover the prospect, every adverse party, related companies such as parents, subsidiaries and owners, and spouses where family matters are involved. It should run against current clients, former clients, adverse parties in past matters and prospective clients the firm declined.

Prospective clients need care. Model Rule 1.18 addresses information from people who qualify as prospective clients, even when no representation follows, and Opinion 512 notes that duty in its discussion of AI tools. Your jurisdiction's version of the rule governs, so design intake safeguards around it. Declined prospects belong in the conflict database, and what they told the firm stays inside its systems with access limited to the people who need it.

Name matching is where an agent adds the most. People misspell company names or leave out the LLC. The agent searches every variant it can find in the inquiry and ranks the hits by how closely they match. Each hit shows the matter it came from, so the lawyer clearing it reads a short list instead of a raw export.

Keep a record of every search. The intake memo lists each name searched and each hit, with the lawyer's decision and its date, and it is saved to the new matter or to the declined prospect's record. If a conflict question comes up years later, the firm can show what it checked and who cleared it.

What to automate and what to keep with a lawyer

The split is easy to state. The agent reads, extracts, searches and drafts. A lawyer clears conflicts and decides whether to take the matter and on what terms. Intake staff release administrative replies to the prospect from templates a lawyer approved.

Rules for an AI intake assistant

If the firm puts a chat assistant on its website, the Florida Bar's Ethics Opinion 24-1 says lawyers must inform prospective clients that they are communicating with an AI program, and the lawyer remains responsible for what the chatbot says under the advertising rules. Florida is one state, but telling people they are talking to software is a sensible rule anywhere.

The State Bar of California's 2026 Practical Guidance names autonomously facilitating client intake as an example of agentic AI. It says attorneys must not deploy agentic systems in a way that lets them make substantive legal determinations or communicate legal advice without meaningful attorney supervision and review. Our proposed design reserves legal advice and engagement decisions for a lawyer, and intake staff may release approved administrative acknowledgments.

Speed still matters inside those limits. With the reply already drafted and the conflict results attached, the person releasing it reviews a draft instead of starting from a blank email. The facts of the case and any opinion on it wait for a lawyer.

Engagement letters and AI disclosures

Opinion 512 calls the engagement agreement a logical place for disclosures about AI use. It also says that general, boilerplate provisions in an engagement letter that purport to authorize AI use are not sufficient for informed consent. If your firm will put client information into an AI tool in a way that needs consent, the letter has to name what is being used and why, in terms the client can understand.

The agent drafts engagement and non-engagement letters from your templates, with the parties and the scope filled in from intake. A lawyer reviews every letter before it goes out. Non-engagement letters deserve the most care, because what they say about deadlines and next steps can matter later.

Working inside your practice management system

Many firms run intake through a practice management system or an intake add-on. Clio's 2024 report lists Clio Grow for client intake and CRM next to Clio Manage for practice management. PracticePanther lists client intake among its practice management features, and Filevine sells Leads and Client Intake as part of its platform.

The agent sits in front of whichever one you run. It reads the inquiry from the shared inbox or the web form and runs the conflict search through the system's own data. Then it writes the new contact and the intake memo back to it. The sprint checks how your system exposes contacts and matters before we quote a build, and you can start from email and web forms alone, which needs no system access during the sprint.

Whatever the system, the agent writes to it through an account with only the permissions intake needs, and every record it creates is marked as drafted by the agent until a person approves it.

Worked example: what the hours look like

This is an illustrative example with assumed numbers, not a client result. Take a 30-person firm with a mixed injury and business practice that gets 60 new inquiries a week across the web form and the phones. Today each inquiry takes about 20 minutes on average to read, reply, enter in the system and run conflicts for those that go to a consult. Sixty inquiries at 20 minutes is 1,200 minutes, or 20 staff hours a week.

Assume 45 of the 60 are routine. With the agent, intake staff spend about 5 minutes reviewing each drafted reply and its conflict results, 225 minutes in all. The other 15, such as those with unclear facts or possible conflicts, still take the full 20 minutes, another 300. The total falls to 525 minutes, about 8.8 hours, which gives back about 11.3 staff hours a week.

Speed is the measure that matters more than hours. Track the share of inquiries answered within one business hour and on the same day, against your own baseline from before the agent. Clio's benchmark is a useful comparison, but your own phones and inbox are the test that counts.

Starting small with intake

Count one ordinary week. Have intake staff note every inquiry, the minutes it took, whether it went to a consult and how long the prospect waited for a first reply. Then run an agent on copies of last month's inquiries, with nothing going to prospects, and compare its conflict results with the ones your team ran.

That is roughly what our five-day AI Workflow Sprint does, for $2,500, and the fee comes off the build. Full refund if we do not find 10 staff hours a week. A build is from $7,500 per workflow and running it is $1,200 a month.

Common questions