Sorting Client Enquiries: AI Chat for Law Firms

In a law firm, a lot is decided at the very first contact. Anyone seeking advice on a legal problem is often tense, under time pressure and rarely turns to just one address. If the phone rings out or a message goes unanswered for days, the inquiry moves on to the next firm. At the same time, intake almost everywhere runs by hand: the front desk notes the matter on a slip of paper, asks about the area of law and the urgency, and tries to arrange a callback. That the first impression is increasingly digital is shown by the figures: the internet is today the most frequently used source of information when searching for a lawyer, and for more than half of prospective clients a well-made website is decisive in engaging a lawyer (Soldan Institut). This is exactly where an AI chat assistant comes in. It captures first inquiries around the clock, sorts the area of law and the matter in a structured way and initiates an appointment when needed, without giving legal advice itself. This article shows which tasks a law firm assistant takes on, where the limit to legal advice under German legal services law lies, how data protection and attorney confidentiality are preserved, and how the front desk can be noticeably relieved without anyone giving up personal contact.

Enlarge graphic

Labels in the illustration

• My employer has terminated
• my contract — what can I do?
• I sort your matter
• (employment law) and take
• the key details — no
• legal advice.
• Available around the clock
• XICBOT
• sorts ·
• routes on
• Structured intake sheet
• Legal area
• Employment law
• Matter
• Review dismissal
• Urgency
• Deadline: 3 weeks
• Contact
• email on file
• Proposed slot
• Thu 10:00
• No legal advice: the assistant sorts inquiries and hands them to the firm — RDG-compliant
• 60%
• of lawyers won clients in 2017
• via online portals (Soldan Institut)
• 86%
• satisfied with human service
• in online shopping (Bitkom)
• 24/7
• capture inquiries, after
• hours too (project experience)
• Relieves the front desk — structured inquiries, not sticky notes

Structure and sort incoming client inquiries

Why the First Client Inquiry Decides the Engagement

The market for legal services is densely occupied. As of 1 January 2026, 167,547 (Bundesrechtsanwaltskammer) lawyers were admitted in Germany, an all-time high. Anyone searching therefore has a choice and compares. What matters is thus not only whether a firm fits professionally, but also whether it is reachable at the moment of need and responds quickly. The first channel is almost always digital: the internet is the most frequently used source of information in a lawyer search, and as early as 2017, more than 60 percent (Soldan Institut) of lawyers won new engagements through online lawyer-search services and referral portals. Those who are found online but do not answer at the decisive moment give away exactly the inquiries they previously won through visibility.

At the same time, the behaviour of those seeking advice has shifted. A large share of website visits now comes from mobile devices, that is from people who want to clarify quickly on the go whether they are in the right place. A considerable share of these first contacts arises outside office hours, in the evening after work or at the weekend (project experience), when no one is reachable at the front desk. A chat assistant replaces neither the phone nor the personal conversation but complements the serial, hours-bound route with a parallel, round-the-clock channel through which an inquiry is captured in a structured way instead of vanishing in the busy signal. Much as a shop assistant recovers abandoned carts in chat, a law firm assistant catches first inquiries that would otherwise be lost.

What the Assistant at the Front Desk Takes On

A large part of the first communication is about recurring, well-structurable things: which area of law it concerns, how urgent the matter is, whether there is a deadline, which documents exist and how the person can be reached. This intake can be reliably moved into the chat, so the front desk spends less time noting and sorting and more time on the matters already in progress. What matters is the clear division of roles: the assistant handles the orderly routine of intake, the legal work stays entirely with the team. A detailed overview is on the page about the AI assistant for law firms.

The Clear Limit: No Legal Advice (RDG)

The most important point first: a law firm assistant gives no legal advice. German legal services law reserves advice in a specific individual case for the persons authorised to do so, primarily lawyers. An assistant that sorts and forwards stays outside this limit as long as it does not examine the matter legally, assesses no prospects of success and gives no recommendation for action in the specific case. This limit is not a technical weakness but a deliberate construction: the assistant captures, sorts into an area of law and initiates an appointment; the legal appraisal begins only in the conversation with the lawyer. Just as important is honesty about its own abilities: an AI assistant can be wrong, which is why a properly set-up assistant binds its answers to your own, checked content and passes uncertain matters to people rather than inventing something. How free answers can be bound technically to a checked knowledge base is described in the article on preventing hallucinations.

• No assessment of the legal position and no examination of a case's prospects of success
• No recommendation for action or advice on the specific matter of the person seeking help
• No interpretation of contracts, official notices or pleadings as legal advice
• General, publicly available information remains possible, advice in the individual case does not
• For anything that requires legal appraisal, the assistant hands over to the firm

A Client Inquiry Step by Step

An example makes what this feels like in everyday use tangible. Shortly after eight in the evening someone opens the firm website and types into the chat: My employer has terminated my contract, what can I do? The assistant understands this perfectly ordinary description without the person having to click through a menu. It sorts the matter into employment law, points out politely that it gives no legal advice, and captures the key details in a structured way. This ability to understand natural language and combine several pieces of information in the same window sets guided intake apart from a rigid form, as the article on first inquiries via chat instead of a long form shows.

• A greeting that suits the page and the clear note that no legal advice is given
• Capture the matter in normal language and sort it into the fitting area of law
• Actively ask about possible deadlines and flag urgent cases as pressing
• Capture the necessary key details: opposing party, history so far, documents on hand
• Store a contact route and, when needed, initiate an appointment for an initial consultation
• Hand the structured intake sheet to the responsible person in the firm

The difference from a classic contact form lies in the in-between moments. If the person asks midway whether an initial consultation costs anything or how quickly there will be a response, the assistant answers from your stored details and then leads back to the intake. In the end there is no loose note in the inbox but an orderly case with area of law, matter, deadline, contact and a proposed slot. How an appointment request becomes a confirmed appointment is shown in the article on appointment requests via chat assistant and on the page about the booking assistant.

Sorting the Area of Law and the Matter Correctly

A firm is rarely responsible for everything, and larger units split into departments. For an inquiry to reach the right person, the assistant has to do more than capture; it also has to sort. From the description it recognises whether it concerns employment, family, tenancy, traffic or another area of law and forwards it to the responsible contact or the fitting specialist lawyer. That specialisation increasingly matters is shown by the statistic: there were 47,436 (Bundesrechtsanwaltskammer) specialist lawyers as of 1 January 2026, an increase on the previous year. The more clearly an inquiry is assigned to an area of law, the faster it reaches the right department instead of being passed around internally several times. What carries this sorting is well-kept, understandable content, as described on the knowledge base page.

This binding to your own sources is the decisive difference from a generic chatbot. The assistant is instructed not to invent responsibilities and not to promise anything the firm cannot deliver. Delicate checks such as the one for a possible conflict of interest stay expressly with the team; the assistant only prepares them with clean details. And from the guided conversations the firm learns along the way: which areas of law are asked about especially often, where those seeking advice hesitate and which questions recur is made visible by the conversation analytics, so the intake can be sharpened continuously.

Data Protection, Confidentiality and Labelling

In a law firm the demands on confidentiality are particularly high. The mere fact that someone contacts a specific firm with a specific matter is sensitive, and attorney confidentiality already applies in the pre-engagement relationship. That is why data protection belongs in the design from the start, not as an afterthought. A properly set-up assistant processes only the details needed for intake and asks for no details that belong in a confidential conversation with the lawyer. At XICBOT, hosting is in Germany and the answers are generated by a specialised model provider, also outside the EU on the basis of standard contractual clauses; there is a data processing agreement, data sovereignty and a clear deletion concept.

Two further points belong here. First, it must be recognisable that one is writing to an AI and not a human. The EU AI Act provides for such transparency for AI systems in direct contact, which the article on the labelling duty under the EU AI Act goes into in more detail. Second, where a technical connection to the calendar or firm software is set up in the Custom package, it is created through clearly defined actions, often called function calling: the assistant may only call defined functions, such as reading free times or creating a case, and receives no blanket full access. How this controlled interplay is built is described on the page about tool control. The contractual and technical foundations are summarised on the page about privacy and hosting, and how an assistant can be operated in a GDPR-compliant way overall is deepened in the article on data protection and hosting of an AI chatbot.

Phone, Form and Chat Assistant in Comparison

An assistant does not replace every other route to the firm across the board; it fills a gap. The phone stays important but is serial and bound to office hours; a form captures but sorts nothing and answers nothing. The AI assistant sits in between: reachable around the clock, sorting already in the conversation and handing over to a human as soon as it becomes legal. An indication of why the handover to people matters comes from satisfaction with customer service in online shopping: of those who had contact with a human agent there, 86 percent (Bitkom) are satisfied, compared with 50 percent (Bitkom) of those who used a chatbot service. An assistant that takes on intake and hands the rest to people combines both.

In a Few Steps to Your Own Firm Assistant

The path begins with a short look at your most frequent first inquiries, your areas of law and the topics that tie up the front desk the most. On this basis the assistant is trained on your content, embedded into the website with a short snippet and, if desired, connected to the calendar. A fixed rate cannot seriously be promised, because how many inquiries the assistant captures in a structured way depends on the firm, its focus and the website. What is dependable, however, is the mechanism: reachable round the clock, capture in seconds, a clear limit to legal advice and a protected handover to the team. For context: Gartner expects agentic AI to resolve around 80 percent (Gartner) of common service issues without human intervention by 2029 and to reduce operational costs by about 30 percent (Gartner), but at the same time calls for a sober view, because more than 40 percent (Gartner) of agentic AI projects are likely to be cancelled by the end of 2027 due to escalating costs, unclear business value or inadequate risk controls. An overview of functions and packages is given by the services; anyone who wants to run this through their own case can request a free demo for your firm. For plain inquiry intake, the lead assistant bundles the matching scope of functions.

• Gather the most frequent first inquiries and the areas of law to distinguish
• Define the limit to legal advice clearly: sort and forward, do not evaluate
• Bind answers to checked firm content for dependable organisational information
• Set the handover to the responsible person for anything legal and sensitive
• Secure confidentiality, data protection and hosting in Germany contractually
• Label the assistant as AI and keep sharpening via the conversation analytics

Frequent questions