AI intake for personal injury firms that answers at 11pm.

It’s 6:40 on a Friday and somebody just got rear-ended on the interstate. Two hours later they’re sitting in an ER waiting room with a phone in one hand, searching for an attorney.

They call the first three firms on the page. Two roll to voicemail. The third picks up. On Monday morning your intake coordinator returns the voicemail and finds out the caller signed with someone else on Saturday afternoon.

Nothing about that is a marketing problem. The ad spend worked. The lead arrived. It arrived at 8:41 on a Friday night, which is when collisions actually happen, and the intake process that runs fine at 10am on a Tuesday does not exist at 8:41 on a Friday night.

The profession-wide numbers are not flattering. In a 2024 secret-shopper study, a third-party research firm contacted 500 US law firms posing as prospective clients. Only 40% answered the phone, down from 56% in 2019. Only 33% replied to an email, down from 40%. Nearly half could not be reached by phone at all. That is the field you are competing in, which is also the opportunity in it.

One honest note about the 21x statistic

If you have sat through an intake pitch, you have seen a slide claiming that responding within five minutes makes you 21 times more likely to qualify a lead, usually credited to MIT or Harvard. We are not going to use that number, because the attribution is wrong. It traces to a 2007 analysis by Dr. James Oldroyd of platform data from a sales-software vendor, covering six companies. It was never an MIT publication, and the widely repeated line about 78% of buyers hiring the first responder has no published methodology behind it at all.

What is real is the published work: the 2011 Harvard Business Review audit of 2,241 US companies found an average first response of 42 hours, 23% of companies never responding at all, and companies that made contact within an hour roughly seven times more likely to qualify the lead. Seven is a smaller number than twenty-one. It is also a number you can check. The provenance on all of this is documented in a public audit of speed-to-lead statistics, and it is worth five minutes before you buy anything in this category.

We wrote the longer version of this argument in speed to lead in personal injury law, including what a structured intake conversation should actually ask.

This is the part most vendors skip, so we will be specific. An intake agent is a fact-collection and routing tool. It occupies the same seat a new intake coordinator occupies on their first week: ask the questions on the sheet, write down the answers, get them to an attorney. Nothing beyond that.

What the rules actually say

ABA Model Rule 1.18 treats a person who consults your firm about a possible representation as a prospective client. The duty not to use or reveal what they told you attaches even when you never take the matter. That is a design constraint on where intake recordings and transcripts live, who can open them, and how long they are kept.

The comment to ABA Model Rule 5.3 covers nonlawyer assistance from outside the firm, and it names internet-based services that store client information as an example. The lawyer has to make reasonable efforts to ensure the outside service behaves in a way compatible with the lawyer’s own obligations. In practice that means you need to be able to answer, in writing, where the data sits and who touches it.

ABA Formal Opinion 512, issued 29 July 2024, is the first ABA ethics guidance on generative AI. It expects a reasonable understanding of the tool’s capabilities and limits, evaluation of confidentiality risk before client information goes in, and firm-level policies governing use. Worth noting alongside it: in Clio’s 2025 Legal Trends Report, 53% of legal professionals said their firm has no AI policy or they were not aware of one (Illinois Supreme Court Commission on Professionalism, Oct 2025). If you are buying intake software this quarter, the policy is part of the purchase.

On disclosure specifically, the Florida Bar got there first. Its Ethics Opinion 24-1, issued January 2024, is direct about it: if a firm uses a chatbot to interact with prospective clients, those people have to be told they are communicating with AI and not a lawyer. Roughly thirty jurisdictions have now published an opinion, policy, or practical-guidance document on generative AI, and they do not all say the same thing. Yours is the one that matters.

We are not lawyers and none of this is legal advice. Your jurisdiction’s rules and your own state bar’s guidance control, and some of that guidance is stricter than the ABA’s. Have your ethics counsel review any intake system before it goes live, including ours. If a vendor discourages that review, that tells you something.

More general questions are answered on the FAQ page, and the build sequence is laid out on how we work. An intake build of this shape usually sits in the Scale tier.