Skip to content

45% of clients now ask AI first. Most firms are invisible there. The fix.

survey

The short answer: demand moved from 6% to 45% in a year while the supply of firms these systems name stayed near zero

A survey of 1,002 US adults found the share using AI tools to find a local business went from 6% to 45% in twelve months. A separate index of more than 350,000 business locations found AI search names about 1.2% of them. For a law firm the leak is two-sided: after-hours intake nobody answers, and absence from the answer a prospective client is now reading instead of a result list. Both are fixable, and neither is fixed by promising a ranking.

The referral question moved. People used to ask a colleague which lawyer to call. Now a lot of them ask a chat window, and the answer arrives with three names in it.

This is an industry analysis, not a client result. Every number below comes from a published study or an open model, and it carries the label next to it.

Three leaks run through a typical four-lawyer firm. Absence from AI answers, intake lost after 17:00, and replies that arrive after the client has instructed somebody else. Here is what each one costs.

The answer that names three firms, none of them yours

Recommendation behaviour changed faster than most marketing budgets did.

6% → 45%

Consumers asking AI for local recommendations, in twelve months.survey

The supply side has not caught up, which is the opportunity.

1.2%

Of local businesses are named by ChatGPT.study

Being absent is the default state, not a penalty for doing something wrong. Almost every firm in your city is absent too. The one that fixes its data trail first gets named first.

Review signals feed the same machinery.

47%

Of consumers will not use a business with fewer than 20 reviews.survey

Visibility here is measurable, which is the part most firms do not expect. You run real prompts, you count how often the firm is named, and you re-run them on a schedule. Base covers 30 prompt checks a month (product spec). Named or not named. No mystique.

17:05, urgent matter, mailbox

People do not call a lawyer casually. They call after a dismissal letter, an accident, or a deadline they just noticed. That call happens when they read the letter, not when your office is staffed.

The mandate goes to whoever picked up. Not to the best firm. To the reachable one.

The machine answers in seconds and behaves like intake, not like advice. It discloses that it is a digital assistant, takes the matter type, the parties, the urgency and a number, offers a slot, and keeps a human path open at all times. It books. It does not advise.

The reply that came on Thursday

Enquiry Monday, conflict check Wednesday, reply Thursday. Every step is defensible on its own. Together they are the reason the file went elsewhere.

~60×

Less likely to reach a live conversation when you reply next day instead of within the hour.HBR audit, 2,241 firms, 1.25M leads

The drop starts inside hour one. First-hour replies are ~7× more likely to qualify the lead than second-hour replies (HBR audit), and 23% of audited firms never replied at all (HBR audit).

A fast acknowledgement is not a fast opinion. The machine confirms receipt, asks the two or three questions your intake needs, and offers a slot. The legal judgement stays exactly where it belongs.

Put a mandate value on it. The bands go to €5,000.

What does replying late cost you?

≈ €16,250 a month walks away.

Estimate. Model: an audit of 2,241 firms (HBR) found replies after an hour are dramatically less likely to reach a live conversation, next day is ~60× worse. Your real numbers replace it on the call.

Three machines close these three leaks. Install them in this order. All twenty are on the solutions hub.

Front Desk 24/7

Front Desk 24/7 answers the 17:05 call. Every call picked up in seconds, structured intake taken, urgent matters routed to a human immediately. Evenings and weekends included, because that is when people finally have time to call.

Lightning Reply

Lightning Reply gets the first response out while the enquiry is warm. Typical machine first reply: 38 seconds (product spec), on forms, email and portal enquiries, around the clock.

Visible

Visible puts the firm into the answer. A measured baseline with sample sizes, the data trail fixed where AI actually looks, review signals engineered honestly, and re-tests on schedule so drift gets caught early.

MODELLED EXAMPLE

Employment law firm, 4 lawyers, Ruhr area

Clients now ask AI before they ask colleagues. Recommendation queries moved from 6% to 45% of consumers in a year, and AI answers name almost nobody local. A measured baseline, structural fixes, and scheduled re-tests move a firm from absent to cited.

6% → 45%

consumers asking AI

survey

1.2%

of local businesses named by ChatGPT

study

30

prompt checks per month at Base

product spec

The Law Firm Fix List

intake to conflict checks: 12 processes, priced.

PDF · 1 page · email required

Is this data from your clients?

No, and it says so on every number: published studies plus open models, labelled. Client cases join as they mature, with permission and names.

My law firm is different.

The bands are editable in the calculator and the discovery call runs your real numbers. The model is the start of the conversation, never the end.

What would you install first in a law firm?

The three machines above, in that order. The analysis confirms or corrects it for your case in 3 minutes.

Clients ask first, then they instruct.

Three minutes tells you where your firm is missing from the answer.

Run the free analysis

free · 3 minutes · report by mail