AI Agents for Legal and Professional Services: Client Intake Without the Friction
Direct answer: An AI agent for legal and professional services, most valuable to firms as an after-hours legal intake AI agent, handles client intake, answers general legal information questions, explains process steps, and books appointments with solicitors or advisers, 24 hours, in the client's language, without a partner needing to be available. It does not provide legal advice. It does not create a solicitor-client relationship. It handles the layer that keeps both clients and fee earners from reaching each other as efficiently as they should. Ojin, a Berlin company running EU-based infrastructure, supplies the agent layer for this and acts as the firm's data processor rather than as an adviser to its clients.
Law firms have an intake problem. Most of them know it. A potential client with an urgent employment matter, a family dispute, or a time-sensitive property transaction cannot always wait until 9am on a weekday to speak to someone who can determine whether the firm can help. They search online at 8pm. They find the firm's website. They see a "contact us" form and a phone number that goes to voicemail after hours.
Some fill in the form. Many do not. The firm learns about this as a revenue problem, a gap between traffic and enquiries, without a clear mechanism for what changed it. A Human AI agent that handles the first conversation at any hour is the mechanism.
Where is the line between legal information and legal advice?
The line is simple: an AI intake agent can explain legal information and process, while anything that applies the law to a client's specific situation requires a qualified solicitor. That distinction exists because the solicitor-client relationship, regulated by the SRA in the UK and equivalent bodies across the EU, creates legal obligations an AI agent must not inadvertently trigger, and advice specific to a client's situation risks professional liability and regulatory exposure.
For example, an AI agent that tells a client "based on what you've told me, you have a strong case" has crossed from information into advice. An Ojin deployment holds that line in the knowledge base rather than in the prompt alone, because a boundary a firm is regulated against should not depend on how a client phrases the question.
Information the agent can provide: what a legal process involves, what documents are typically needed, what the timeline looks like, what the firm's areas of practice cover, and what the next step is in engaging the firm. None of this constitutes legal advice. All of it is genuinely useful to a prospective client deciding whether to proceed.
The advice line: specific opinions on a client's legal position, prospects, or strategy, these require a qualified solicitor. The agent escalates all advice-adjacent questions: "That is a question I want to make sure you get the right answer to, let me connect you with one of our team."
The Solicitors Regulation Authority has not set AI-specific rules on this exact point. Its Risk Outlook report on AI in the legal market treats the boundary as already covered by existing conduct obligations, holding the named solicitor accountable for any output regardless of the tool that produced it, and the Law Society's guidance on generative AI takes the same outcome-focused approach, treating human oversight of anything resembling advice as non-negotiable (SRA; Law Society).
What does an AI agent handle during legal client intake?
An Ojin agent in legal intake handles three core tasks: identifying the new client and their matter, routing them to the correct practice area, and confirming the documents needed before the first consultation. Each of these removes a point of friction that would otherwise sit with a receptionist or a fee earner.
New client identification. The first call with a prospective legal client has a standard structure: who they are, what the matter involves, which practice area it falls under, urgency level, and whether the firm has a conflict. An AI agent can collect all of this information conversationally, structure it for the intake team, and confirm whether an initial consultation has been booked.
Practice area routing. A client who has been in a car accident does not know whether they need a personal injury solicitor, an insurance specialist, or a criminal defence lawyer (if the accident involved a prosecution). An AI agent that asks the right clarifying questions and routes to the correct practice area removes a significant source of frustration from the intake process.
Document checklist. Most legal matters require documents at the outset, identity documents, contracts, correspondence, evidence. An AI agent that explains what to bring to the first consultation and why reduces the first-meeting friction and speeds up matter opening.
Why does after-hours legal intake matter so much?
After-hours availability matters disproportionately because legal problems do not wait for office hours: a prospective client with an urgent matter who cannot reach a solicitor until Monday will often contact another firm first. An AI agent that is available at any hour, collects the intake, and books the next available consultation, captures that client before they leave.
This is the specific gap Ojin's Human Agents are aimed at, the hours when a firm's phone is answered by nobody and its competitor's is answered by something. An employment matter discovered on a Friday afternoon, a dismissal letter, a disciplinary notice, is a typical example: a Human AI agent available that evening collects the initial intake information and books a Monday morning consultation, converting the prospect before they search for another firm over the weekend.
The same applies to family law (often disclosed at weekends), immigration (deadlines do not respect office hours), and criminal law (which explicitly requires out-of-hours availability). The AI agent is not a replacement for the duty solicitor. It is the bridge to the appropriate human at the first available moment.
Clio's Legal Trends Report data points in the same direction: a large share of after-hours and weekend contact tends to come from people who have never called the firm before, first-time enquiries rather than existing clients with routine questions, which is exactly the population most likely to instruct whoever answers first (Clio).
How do GDPR and legal professional privilege apply to AI intake data?
Legal intake data is handled under a GDPR-compliant framework, with explicit consent, purpose limitation, and retention limits set out in a Data Processing Agreement between the firm and Ojin, and it does not automatically attract legal professional privilege, since an AI agent is not a solicitor and privilege applies only to solicitor-client communications.
Where that data physically sits is part of the same question. Ojin is a Berlin company running EU-based infrastructure, so a UK or European firm is not adding a cross-border transfer assessment to an already sensitive file. Legal data is sensitive by definition. Matters involving litigation, family breakdown, employment disputes, or criminal proceedings involve some of the most personal information a person shares with any professional.
The design implication: the agent collects process information (identity, matter type, document list) rather than substantive legal information (what actually happened, the client's version of events) in the pre-instruction phase. Substantive instructions take place after the solicitor-client relationship is formally established.
As the ICO's guidance on lawful basis sets out, a firm must identify and document which of the six Article 6 UK GDPR grounds it is relying on before it collects a prospective client's details, and be able to justify that choice on request (ICO).
A firm does not need to procure a platform to answer the question this article opens with. The Ojin platform is self-serve, so a practice can build an intake agent scoped to information rather than advice, and see how it handles its own enquiry types before committing to anything. That model is described in what a self-serve AI platform is.
Frequently asked questions
Does the AI intake agent create a solicitor-client relationship?
No. The intake agent collects information and books appointments. The solicitor-client relationship is created when the solicitor issues a client care letter and the client confirms engagement. The AI agent explicitly communicates this at the start of the interaction.
How does the firm prevent the agent from inadvertently giving legal advice?
The knowledge base is scoped to process information and general legal information, what the law says, not how it applies to this client's situation. The agent is configured with advice-adjacent question detection: when a question approaches the "apply this to my situation" territory, it escalates. The configuration and testing of this boundary is part of the deployment process.
Can the agent handle conflicts of interest checks?
It can collect the information needed for a conflicts check (opposing party name, matter type, connected persons) and flag the data for the firm's conflicts team. It cannot perform the conflicts analysis, that requires access to the firm's matter management system and human review.
For what a deployment like this has to satisfy on data protection, see GDPR and AI agent deployments, and for the reputational guardrails, brand safety when deploying a Human AI Agent. For the category, see what a Human AI Agent is and how it actually works, what a Human AI Agent is and when it beats recorded video, and what a conversational AI agent is and how it differs from a chatbot. Client intake, after-hours cover, accountancy and advisory work, the attorney-client relationship question, and the intake ROI figures each get their own article in this vertical. Across the rest of this batch, the closest neighbours are financial services, where the same advice boundary applies, and HR and employee onboarding, the employer-side version of the same intake conversation.
Ojin legal and professional services deployment: ojin.ai/enterprise. Try a live Ojin agent: docs.ojin.ai.