You're paying for leads your intake team never talks to.
Not because the leads are bad. Because nobody picked up, nobody called back, or somebody decided a good case was a bad one.
AI can fix the speed problem in 2027. It can also make the judgment problem worse.
This guide shows you where to point it. Here’s what to fix before 2027:
- Answer every lead in seconds, on every channel, 24/7.
- Write your qualification rubric before you buy any software.
- Qualify the AI-informed lead who already asked ChatGPT about their case.
- Disclose the bot before Colorado’s law kicks in on January 1, 2027.
- Fix your text and callback follow-up for the new TCPA opt-out rules.
- Keep a human on every yes and every maybe.
- Audit what the AI rejects, every week.
- Measure it in cost per signed case, not chats handled.
Get #2 wrong and the other seven don’t matter.
What Is AI Lead Qualification for Law Firms?
- Step 1CaptureAI asks every lead your intake questions, 24/7.
- Step 2ScoreAnswers are checked against your rubric.
- Step 3RouteQualified leads reach a human in minutes.
- Step 4Attorney decidesAI never says no alone. Your attorney signs.
AI lead qualification is software that collects a prospective client’s details, scores the lead against your firm’s criteria, and routes it to the right person in seconds.
It doesn’t decide which cases you take. Your attorneys do that.
It runs on three parts:
- Capture: an AI chat, voice, or text agent asks the intake questions, any time of day.
- Score: the system compares the answers to your rules (case type, injury, fault, timing, location).
- Route: hot leads go to a human immediately. Everything else gets a defined next step.
Here’s the short version: AI doesn’t qualify leads. Your criteria do.
AI just applies them faster and more consistently than a tired intake specialist at 9 PM.
Why Intake Is Still Where Law Firms Lose Cases
Out of every 10 calls to a law firm, 6 went unanswered.
And it got worse since 2019 2019 2024
Most firms think they have a lead problem. Most have a response problem.
According to Clio’s 2024 Legal Trends Report, only 33% of the 500 law firms its secret shoppers emailed responded, down from 40% in 2019.
Phones were no better. ABA Law Practice magazine reported that only 40% of firms in the same study answered phone calls, down from 56% in 2019.
Read that again: 6 in 10 calls went unanswered.
If you’re spending on PPC for law firms, every one of those missed calls is a click you already paid for.
Speed: Leads Go Cold in Minutes
The best data on response time is old, but nothing since has contradicted it.
Harvard Business Review’s 2011 research analyzed 1.25 million sales leads at 42 U.S. companies. Firms that tried to contact a lead within an hour were nearly 7 times as likely to qualify it as firms that waited even one hour longer, and more than 60 times as likely as firms that waited 24 hours or more.
A separate audit in the same research tested 2,241 companies. Among those that responded at all, the average response time was 42 hours.
A personal injury prospect doesn’t wait 42 hours. They call the next firm on the page.
Consistency: Every Intake Specialist Has Bad Days
Human-only intake
- Skips questions when the queue backs up
- Screening changes by who answers
- Voicemail after hours
AI-assisted intake
- Same questions, same order, every time
- Every answer logged in your CRM
- A human still makes the call
Speed isn’t the only leak. Judgment is.
Your best intake specialist asks every question. Your newest one skips three when the queue backs up.
- ❌ The Friday 6 PM caller gets a rushed screen.
- ❌ The Spanish-speaking caller gets a voicemail.
- ❌ The “maybe” case gets marked “no” because nobody wanted to dig.
- ✅ AI asks the same questions in the same order, every time.
- ✅ It documents every answer in your CRM.
- ✅ A human still makes the decision.
That last line matters more than anything else in this article.
The 8 Priorities for Law Firm Intake in 2027
8 priorities, 5 jobs. Work from top to bottom.
Most firms will spend 2027 shopping for AI tools.
The ones that sign more cases will fix what the AI is told to do first.
Here are the eight priorities, in the order you should tackle them.
1. Answer Every Lead in Seconds, on Every Channel
Six channels. One queue. One clock.
Your response time is only as fast as your slowest channel.
Most firms track phone calls. Few track how long a Facebook lead form sits before anyone touches it.
Map every place a lead can come in:
| Channel | Where leads usually die | What AI can do |
|---|---|---|
| Phone calls | After hours, lunch, overflow | AI voice agent answers, captures facts, books a callback |
| Website forms | Sits in an inbox | Instant text or call within seconds of submission |
| Website chat | No one staffing it | AI chat runs the intake questions 24/7 |
| Google LSA calls and messages | Missed calls hurt your LSA standing | AI answers and logs every call |
| Facebook lead forms | Hours of delay before first touch | Instant text and call trigger |
| Inbound texts | Personal phones, no logging | AI replies and pushes the lead to your CRM |
Set one rule: every lead gets a first response in under five minutes, 24/7.
AI makes that rule possible without hiring a night shift.
When WEBRIS built Facebook campaigns for Shook & Stone, one-tap forms and faster intake calls drove higher contact rates. The channel only works if intake keeps up.
2. Write Your Qualification Rubric Before You Buy Software
| Criterion | Qualifies | Disqualifies | Send to attorney review |
|---|---|---|---|
| Case type | Matches your practice areas | A practice area you don’t handle | Edge cases you rarely take |
| Injury | Injury with medical treatment | No injury, no treatment | Delayed treatment |
| Fault | Other party likely at fault | Caller clearly at fault | Shared or disputed fault |
| Timing | Well inside the filing deadline | Deadline clearly passed | Close to the deadline |
| Location | Your jurisdiction | Outside every state you serve | Incident in one state, client in another |
| Representation | No current attorney | Already represented | Wants to switch attorneys |
This is where most AI intake projects fail.
The vendor asks, “What makes a qualified lead?” and the firm says, “You know, a good case.”
AI can’t apply criteria you never wrote down.
Start with your personal injury intake form and turn it into rules like the table above.
Notice the last column.
Every “maybe” goes to a human. The AI never gets to turn a gray-area case into a “no.”
Have an attorney approve the rubric before it goes live.
3. Qualify the AI-Informed Lead
The lead you trained for
- Knows little about their case
- Expects you to explain everything
- Calls the first firm they find
The 2027 lead
- Already asked an AI assistant
- Arrives with assumptions about fault, value and deadlines
- May have been told they don’t have a case
Your 2027 leads talk to AI before they talk to you.
According to Clio’s 2026 Legal Trends Report, nearly 1 in 4 U.S. adults has used AI to answer a legal question, and 46% say they would consult AI before contacting a lawyer.
That changes intake in two ways.
First, some leads arrive with wrong assumptions about fault, case value, or deadlines. Your intake script needs a question like: “Have you looked into this anywhere else yet? What did you find?”
Second, some good cases never call because the AI told them they didn’t have one. That’s a marketing problem, not an intake problem.
It’s why ranking in ChatGPT and generative engine optimization now sit upstream of intake.
Your AI intake agent should never argue with the lead’s research. It should collect the facts and let your attorney sort out the rest.
4. Disclose the Bot
| State | Law | What it requires | Effective |
|---|---|---|---|
| Maine | 10 M.R.S. §1500-DD | No chatbot use that could mislead a reasonable consumer into thinking they’re talking to a human, unless you clearly disclose they aren’t | In force since 2025 |
| Utah | Utah Code §13-77-103 | Disclose AI when a consumer clearly asks. Some licensed professions must disclose up front in high-risk interactions | In force |
| Colorado | HB 26-1263 | Public conversational AI must disclose it’s AI, and may not imply its output comes from a licensed professional | Jan 1, 2027 |
| California | AB 1609 | Customer service chatbot rules, including a path to a human | Jan 1, 2027 ($500M+ revenue businesses) |
This is general information, not legal advice. Rules vary by state and by tool, so confirm how they apply to your firm with your own counsel.
This is the biggest change for 2027.
If your AI chat or voice agent sounds human, several states now require you to say it isn’t.
Two of these matter most for law firms.
Colorado’s law bars a bot from implying its answers come from a licensed professional. Your intake agent should never sound like it’s giving legal advice from an attorney.
The Colorado law has exclusions for some business tools, so have counsel confirm whether your intake bot is covered. Disclose either way.
Maine’s rule doesn’t require anyone to be fooled. Using a bot in a way that could mislead is enough.
Utah’s up-front rule covers occupations regulated by the state Department of Commerce. Utah’s practice of law is regulated under the Utah Supreme Court’s rules, so confirm with counsel whether the up-front rule reaches your firm.
The safest move is the same either way. And it helps conversion.
In Clio’s 2026 report, 82% of consumers said they want to know when their lawyer uses AI, and 70% said they trust a lawyer more when that use is disclosed.
- ✅ Open every AI chat and call with a plain line: “I’m the firm’s AI intake assistant. I’ll collect a few details and get you to our team.”
- ✅ Answer honestly every time someone asks, “Am I talking to a real person?”
- ✅ Give every lead a fast path to a human.
These laws are changing fast. Check every state where you take cases, not just where your office sits.
5. Rebuild Text and Callback Follow-Up for the New TCPA Rules
1Lead submitted form and agreed to texts, 9:42 PM
- 1Consent firstTexts start only after the lead agrees on your form or call.
- 2Say it’s AIThe first message names the assistant and your firm.
- 3Opt-out in every messageDisclose your designated method, like “Reply STOP.”
- 4Honor it fastOne confirmation, then silence. Log it in your CRM.
AI follow-up is mostly texts and automated calls. Both run through the Telephone Consumer Protection Act.
Two rules matter here.
AI voices count as artificial voices. The FCC’s February 2024 Declaratory Ruling (FCC 24-17) confirmed that AI-generated voices fall under the TCPA’s rules for “artificial or prerecorded” voice calls, including consent requirements.
Opt-out rules changed on November 25, 2026. Under FCC Order 26-67, callers may designate exclusive methods for revoking consent, such as a reply keyword for texts or a key press on calls.
If you designate one and disclose it clearly in the message, you don’t have to process opt-outs made other ways. If you don’t, you must keep honoring opt-outs made by “any reasonable means.”
Here’s what that means for you:
- ✅ Capture consent on every form and call before AI follow-up starts.
- ✅ Put your opt-out method in every automated text and call.
- ✅ Make sure your AI vendor logs and honors opt-outs fast.
- ❌ Don’t let an AI agent keep texting someone who said “stop” in plain English.
Have your counsel review your consent language. TCPA suits are expensive, and plaintiffs’ firms know exactly where to look.
6. Keep a Human on Every Yes and Every Maybe
| Tier | What it means | What happens |
|---|---|---|
| Tier 1: Qualified | Meets every must-have | Human calls within 5 minutes. Attorney reviews same day |
| Tier 2: Maybe | Missing info or a gray-area answer | Human follows up same day to fill gaps |
| Tier 3: Disqualified | Clear disqualifier | Human spot-checks, then sends a clear decline and a referral where appropriate |
AI should sort. Humans should decide.
This isn’t only good practice. It’s what the ethics guidance expects.
ABA Formal Opinion 512, issued July 29, 2024, applies the existing Model Rules to generative AI. That includes competence (Rule 1.1), confidentiality (Rule 1.6), and supervision of lawyers and nonlawyer assistance (Rules 5.1 and 5.3).
In plain terms: you’re responsible for what your AI intake agent says and decides. You need to understand the tool, protect what prospects tell it, and supervise it.
ABA opinions aren’t binding on their own. Check your state bar’s guidance too.
Three hard rules for any AI intake agent:
- ❌ No legal advice. Ever.
- ❌ No promises about case value or outcomes.
- ✅ A clear statement that no attorney-client relationship exists until the firm confirms one.
7. Audit What the AI Rejects, Every Week
The weekly rejection audit
- Pull every lead the AI marked disqualified
- Have a senior intake person or attorney re-read 10 to 20 of them
- Log any that should have been Tier 1 or Tier 2
- Find the rule that misfired and fix it
- Track your false-negative rate week over week
The most expensive AI mistake is the case you never see.
A missed call is visible. A good case quietly scored as “no” is not.
Run the audit above every week. Your false-negative rate is good cases wrongly rejected, divided by all rejections.
If it isn’t trending toward zero, your rubric is the problem, not the AI.
The audit also tells you whether marketing is the issue. If rejections cluster around one source, read why Google Ads sends junk leads before you blame intake.
8. Measure Qualification in Cost per Signed Case
| Metric | Before | After fixing intake |
|---|---|---|
| Cost per lead | $750 | $750 |
| Lead-to-signed rate | 25% | 30% |
| Cost per signed case | $3,000 | $2,500 |
Your AI vendor will report “conversations handled” and “leads qualified.”
Neither one pays your bills. Signed cases do.
Same ad spend. Same clicks. Same leads.
To run this with your own numbers, plan your click cost from the high end of Google Keyword Planner’s top-of-page bid for your keywords, per click, and assume 25% landing page conversion.
A five-point jump in lead-to-signed rate cuts your cost per signed case by $500. On 20 signed cases a month, that’s $10,000 back every month.
Your exact numbers will differ. The point is where to look: fix intake before you raise your budget.
Track these six numbers monthly:
- Median time to first response, by channel
- Contact rate (leads you actually spoke to)
- Qualified rate (Tier 1 and Tier 2 as a share of all leads)
- Lead-to-signed rate
- False-negative rate from your weekly audit
- Cost per signed case
What Not to Do With AI Intake in 2027
Don’t
- Buy the tool before writing the rubric
- Let AI auto-decline with no human review
- Hide that the agent is AI
- Judge the tool by “chats handled”
- Fix intake and ignore the landing page
Do
- Rubric first, approved by an attorney
- Humans decide every tier
- Disclose up front in every state you serve
- Judge it by cost per signed case
- Fix both
Every mistake on the left has the same root: treating AI as the strategy instead of the tool.
On the landing page point, our lawyer landing page guide covers why long forms kill conversions. If you use AI chat, it replaces the long form, not the phone number.
Your 2027 Intake Checklist
Before January 1, 2027
- Every channel mapped, with a response-time target
- Written qualification rubric, approved by an attorney
- AI disclosure line added to every chat and voice flow
- Fast human handoff available on request
- Consent language reviewed by counsel
- Opt-out method designated and disclosed in every automated text and call
- No legal advice or outcome promises anywhere in AI scripts
- Tier routing set, with a human on every tier
- Weekly rejection audit scheduled
- Cost per signed case tracked monthly, by lead source
If your personal injury lead generation is producing volume but not cases, start at the top of this list.
Frequently Asked Questions
The four questions firm owners ask most before turning on AI intake.
Will AI reject good cases?
Every “maybe” goes to a human.
It can, if your rubric is wrong or if you let it auto-decline.
Prevent it by sending every “maybe” to a human and auditing rejected leads weekly.
Does AI intake replace my intake staff?
AI handles
- After-hours and overflow
- Collecting basic facts
Your staff handles
- Calling Tier 1 leads fast
- Signing the case
No. It covers the hours and volume your team can’t.
Your staff spends less time collecting basic facts and more time calling Tier 1 leads and signing them.
Do I have to tell callers they’re talking to AI?
“I’m the firm’s AI intake assistant. I’ll collect a few details and get you to our team.”
In some states, yes, depending on the bot and the state. Colorado’s disclosure duties start January 1, 2027.
Even where it isn’t required, disclose it. Clio’s 2026 data shows most consumers want to know, and most trust lawyers more when they’re told.
Can an AI intake agent give legal advice?
AI must not
- Evaluate the case
- Estimate its value
- Advise on next legal steps
AI can
- Collect the facts
- Explain the firm’s process
- Route the lead to a human
No. It should collect facts, explain next steps in your process, and route the lead.
Anything that sounds like advice or a case evaluation should come from a licensed attorney.