Definition · HR technology and AI in hiring
Conversational AI
Definition
Conversational AI is technology that lets software hold a natural conversation with a person by text or voice: understanding what they say, keeping track of context and replying in plain language. In HR and recruiting it powers chat and voice agents that answer candidate and employee questions, screen applicants, schedule interviews and collect information at any hour.
Also called: Conversational AI recruiting, Voice AI, Voice agents, AI chat and voice assistants
How conversational AI works
A conversational AI system combines a few pieces:
- Language understanding. Working out what the person wants from free text ("can I come Thursday after 3 instead") rather than requiring them to press a button.
- Dialogue management. Remembering what has been said, what is still missing, and what to ask next.
- Response generation. Replying in natural language, today usually with a large language model guided by instructions and approved content.
- Integrations. Reading and writing to the ATS, calendar or HRIS so the conversation can actually do something, such as book a slot or update a status.
- For voice: speech recognition to turn audio into text and text-to-speech to reply, with low enough delay that the call feels natural.
Earlier chatbots were decision trees: fixed buttons and scripted paths. Modern conversational AI handles unexpected phrasing, follow-up questions and topic changes, which is why it is now used for phone calls as well as chat.
Uses in recruiting
In high-volume hiring, the bottleneck is usually recruiter time for repetitive conversations. Conversational AI takes on those conversations, which is the core idea behind conversational recruiting.
- Answering candidate questions about pay, shifts, location and benefits within seconds, day or night.
- Screening with knockout questions (schedule, work authorization, required license) by text or a short voice call.
- Scheduling interviews, rescheduling and sending reminders.
- Structured interviews by voice or video, where an AI interview agent asks the same questions of every candidate and records answers for review.
- Re-engaging past applicants and silver medalists for new openings.
A typical example: a home care agency gets 600 caregiver applications a month, many submitted after 8 p.m. A text-based assistant replies within a minute, asks five screening questions and offers interview slots. Recruiters start the next morning with a calendar of screened, booked candidates instead of a list of people to call.
Uses in HR
For current employees, conversational AI sits behind HR help desks and self-service: answering policy and benefits questions, checking PTO balances, handling simple requests and routing complex cases to the right HR specialist. It also supports learning, for example voice-based role-play where a new contact-center agent practices handling an angry caller before going live.
Conversational AI vs a basic chatbot
Every conversational AI product is a chatbot of sorts, but not every chatbot is conversational AI. A rule-based recruiting chatbot offers buttons and fixed answers. It is predictable and cheap, and fine for simple FAQ and scheduling. Conversational AI understands free-form language and can manage a real back-and-forth, which matters for voice, for candidates who do not fit the script, and for multi-step tasks. The trade-off is that its output is less predictable, so it needs guardrails, testing and monitoring.
Rules and risks
- Disclosure. Tell people they are talking to an AI. Several US states require disclosure in some contexts, and the EU AI Act requires that people be informed when they interact with an AI system.
- Hiring decisions. If the agent screens, scores or rejects candidates, it can be an automated employment decision tool under NYC Local Law 144, a high-risk system under the EU AI Act, and from January 2027 a tool that triggers notice, explanation and human review duties under Colorado's AI law. Keep knockout criteria job-related, monitor pass rates by group, and give candidates a way to reach a person.
- Recorded calls. Voice agents that record need consent where state law requires all-party consent. Illinois's Artificial Intelligence Video Interview Act adds notice and consent rules for AI analysis of video interviews.
- Accessibility. Offer alternatives for candidates with disabilities, such as a human call or text option, as a reasonable accommodation under the ADA.
- Accuracy. Ground answers in approved content, and test regularly for wrong answers about pay, benefits or policy.
Rules vary by state and country.
Frequently asked questions
What is conversational AI in HR?
It is AI that holds natural text or voice conversations with candidates and employees to answer questions, screen applicants, schedule interviews and complete HR requests, usually connected to the ATS or HRIS.
What is a voice agent in recruiting?
It is a conversational AI that talks to candidates by phone, for example to run a short screening call, confirm availability or schedule an interview, using speech recognition and natural-sounding speech.
How is conversational AI different from a chatbot?
Basic chatbots follow fixed scripts and buttons. Conversational AI understands free-form language, keeps context across turns and can handle unexpected questions, which makes it usable for voice and more complex tasks.
Do candidates have to be told they are talking to AI?
Disclosure is best practice everywhere and required in some situations, including under the EU AI Act and certain state laws. Laws on AI in hiring may also require notice that an automated tool is being used.
What is conversational AI recruiting?
It is the use of AI chat and voice agents to handle recruiting conversations at scale, such as answering questions, screening, scheduling and follow-up, so recruiters can spend more time on candidates who are qualified and interested.