AI / Voice
Large Language Model (LLM)
A large language model (LLM) is a machine-learning model trained on large text collections to interpret and generate language from prompts and context.
A large language model (LLM) is a machine-learning model trained on large text collections to interpret and generate language from prompts and context.
A large language model, or LLM, is a machine-learning model trained on large collections of text to interpret and generate language. In a conversational agent, the LLM can help understand an open-ended reply, compose a response, summarize a conversation, or classify what the person wants. It is one component of the system, not the whole voice agentVoice agentA voice agent is a conversational software system that listens, responds, follows a defined workflow, and can take approved actions during a phone conversation..
The LLM's role in a voice workflow
During a phone call, speech-to-text first converts audio into text. The LLM receives that text together with instructions, recent conversation context, and any approved retrieved information. It produces text or a structured decision that the workflowWorkflowA workflow is a defined sequence of steps, decisions, actions, delays, and outcomes used to complete a business process. can use. Text-to-speech then turns the response into audio. Telephony, permissions, integrations, routing, and data retentionData retentionData retention defines how long recordings, transcripts, messages, personal data, and conversation metadata are kept before deletion or archival. sit outside the model itself.
Practical uses
- Interpret natural variations of the same intent.
- Ask a relevant follow-up question based on an earlier answer.
- Create a concise call summary for a CRM record.
- Draft an answer using information retrieved from an approved knowledge base.
An LLM is probabilistic, so it should not be treated as the final authority for critical facts or business rules. Thoughtly's Agent Builder lets teams combine adaptive prompts with deterministic outcomes when a decision must follow exact logic.
Frequently asked questions
Is an LLM the same as conversational AI?
No. An LLM can power language understanding and generation, while conversational AIConversational AIConversational AI is software that interprets natural language and produces context-aware responses through voice, messaging, or chat. also includes state, interfaces, rules, tools, and actions.
Does an LLM hear or speak audio directly?
Not necessarily. Most voice systems use separate speech-to-text and text-to-speech components around the language model.