Glosario

IA para bufetes de abogados, definida claramente.

Los términos que surgen en nuestro trabajo: agentes de voz, RAG, GEO, AEO, LiveKit, LangGraph, troncales SIP. Sin jerga de marketing, sin sopa de acrónimos.

A

AI legal intake

AI legal intake is the use of conversational AI — voice agents or chatbots — to qualify potential law-firm clients, capture structured matter data (matter type, jurisdiction, urgency, prior representation, conflicts), and create matters in a practice management system like Clio without manual re-keying. Done well, it shifts intake from a labor-intensive bottleneck to a 24/7 always-on process.

AI voice agent

An AI voice agent is software that answers phone calls and holds two-way spoken conversations with callers, then takes structured actions — qualifying leads, booking appointments, creating CRM records, transferring to a human — based on what the caller says. Modern agents combine speech recognition (turning audio into text), a large language model (deciding what to say next), and speech synthesis (turning the response back into audible speech), with all three running in real time so the conversation feels natural.

Answer Engine Optimization

Answer Engine Optimization (AEO) is the practice of optimizing content to win direct-answer placements in search engines — featured snippets, knowledge panels, "people also ask" boxes, and Google AI Overviews. Where traditional SEO targets the ten blue links and Generative Engine Optimization targets AI-generated answers in chat interfaces, AEO targets the structured-answer real estate inside the search results page itself.

C

Conversational AI for law firms

Conversational AI for law firms is AI software — voice agents, chatbots, or both — designed for the specific demands of legal client engagement. Unlike generic conversational AI, it includes practice-area-aware intake flows, conflict checking, ethics-aware messaging that respects state-bar advertising rules, and integration with legal-specific systems like Clio. Most production deployments cover phone intake, web chat, and after-hours coverage as a single coordinated layer.

F

Function calling

Function calling (sometimes "tool calling" or "tool use") is the language-model feature that lets an LLM decide when to invoke an external function with structured arguments instead of producing free text. The developer registers a list of available functions (book_appointment, create_clio_matter, search_documents), each with an input schema. The model decides during a conversation which function to call and produces a structured JSON object matching the schema. The function executes; its result is fed back into the conversation; the model continues. Function calling is what turns a plain chatbot into a useful agent.

G

Generative Engine Optimization

Generative Engine Optimization (GEO) is the practice of optimizing a website and brand entity to be cited inside AI-generated answers from large language models — ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews, and Microsoft Copilot. Where traditional SEO targets the ten blue links, GEO targets the synthesized answer the user sees first; in many query categories, that answer is now the entire user experience.

I

Intent classification

Intent classification is the natural-language-understanding task of mapping a user’s utterance to one of a finite, predefined set of intents — "book appointment", "ask pricing", "report a bug", "qualify a personal injury claim". Traditional conversational AI systems were built around intent classifiers as their central abstraction; modern LLM-based agents do this implicitly through the language model’s reasoning, but the concept still matters in production design.

Interactive Voice Response

Interactive Voice Response (IVR) is the traditional "press 1 for sales, press 2 for support" automated phone system — a finite-state routing tree that uses prompts and DTMF (touch-tone) inputs or limited speech recognition to direct callers to the right destination without a live operator. IVRs ruled business phone systems for thirty years; modern AI voice agents are progressively replacing them in contexts where natural conversation matters more than pure routing.

L

LangGraph

LangGraph is an open-source framework, built on top of LangChain, for orchestrating stateful, multi-step LLM agents as directed graphs of nodes (steps) and edges (transitions). It is particularly useful when an agent needs branching logic, persistent state across turns, sub-agents, or conditional control flow — the cases where a simple "prompt the LLM in a loop" pattern breaks down.

LiveKit

LiveKit is an open-source real-time voice and video infrastructure platform. In the AI voice agent ecosystem, it has become the de facto underlying layer for production deployments — handling audio routing, turn detection, interruption handling, and SIP/telephony integration so agent builders can focus on the language model and tool-calling logic above. LiveKit is what powers the voice layer at Vasquez Law Firm in the Hodos360 reference deployment.

LLM agent

An LLM agent is software that uses a large language model as its decision-making layer to perform multi-step tasks autonomously. Unlike a single LLM call (which produces one response), an agent runs in a loop: it observes the current state, chooses an action (often by calling external tools or APIs), observes the result, and repeats — until the task is complete. This pattern is the foundation of modern AI voice agents, autonomous coding assistants, and many "AI workflow" products.

R

Retrieval-Augmented Generation

Retrieval-Augmented Generation (RAG) is a technique that grounds a large language model in your own data. At query time, a retrieval step finds the most relevant documents (or document chunks) from a corpus you control — typically using a vector database — and includes them in the LLM prompt. This dramatically reduces hallucinations and lets the model answer questions about content the base model has never seen, without the cost of retraining.

S

SIP trunk

A SIP trunk is a virtual phone line that uses Session Initiation Protocol (SIP) over an internet connection to carry voice calls between a phone system and the public phone network. SIP trunks replaced traditional analog and ISDN trunks in most business phone deployments and are the standard way modern AI voice agents connect to the public phone network.

V

Vector database

A vector database is a specialized data store optimized for similarity search over high-dimensional vectors — typically the embedding vectors produced by language models. Unlike a traditional database that finds rows matching exact values, a vector database returns the rows whose stored vectors are mathematically closest to a query vector. This is the underlying primitive that makes retrieval-augmented generation (RAG), semantic search, and recommendation systems work at production scale.

W

Webhook

A webhook is an HTTP callback — one application sending a real-time HTTP POST request to another when a defined event happens. Where polling has the receiver repeatedly ask "anything new?", a webhook has the sender push the event the moment it occurs. Webhooks are how most modern SaaS integrations connect, and how most law-firm intake pipelines pass calls, contacts, and matters between CallRail, GoHighLevel, Clio, and AI voice agents.