AI Recruiting Academy › Module 6: AI across the recruiting desk
The short answer
A recruiting chatbot is an assistant on your careers site, in text messages or in chat apps that answers candidate questions, collects basic details and books interviews at any hour. It helps the candidate experience most where candidates would otherwise wait days for a reply, such as high volume hourly hiring.
It hurts the experience when it traps people in loops, pretends to be human or makes decisions it should not. A good chatbot says it is a bot, answers fast and hands anything sensitive to a person.
Why do recruiters use chatbots?
Because silence is the biggest complaint candidates have. Greenhouse’s 2025 hiring report found 63% of candidates in the US, UK and Ireland felt left in the dark after interviews. Its 2024 candidate report found 20% of job seekers had turned down an offer because of a poor interview experience.
Speed separates the best employers. In the 2025 CandE benchmark research, which covered more than 66,000 candidate responses, the top rated companies kept far fewer candidates waiting months for next steps, and all of the top ten in North America used AI recruiting technology.
A recruiter with 13 open roles cannot answer every “did you get my application?” message within an hour. A chatbot can.
What kinds of recruiting chatbots are there?
| Type | How it works | Good for | Weak at |
|---|---|---|---|
| Rules based bot | Follows set questions and buttons you write in advance | Screening questions, FAQs, booking slots | Anything off script |
| Generative AI assistant | Uses a language model to understand free text and reply in plain language | Open questions about the job, the process and the company | Can give confident wrong answers if not limited to your content |
| Text recruiting | Either type, run by SMS or WhatsApp | Hourly and frontline candidates who rarely check email | Needs consent and opt out handling |
Most products now mix the two: fixed steps for screening and booking, and a language model for questions. Module 1, Lesson 3 explains how a chatbot differs from an agent.
What results have employers reported?
The best documented examples come from high volume hiring. Chipotle launched its assistant, Ava Cado, in 2024. It chats with applicants in four languages, answers questions, collects details, schedules interviews and sends offers to candidates the manager has chosen. Chipotle said it expected the tool to cut time to hire by up to 75%.
In a vendor case study, Paradox says its assistant cut interview scheduling at General Motors from about a week to under 30 minutes. Workday completed its purchase of Paradox in October 2025, which tells you how central these assistants have become.
Note what the bot does and does not do in these examples. It handles questions and logistics. The manager chooses who gets the offer.
Do candidates trust recruiting chatbots?
They trust fast answers more than they trust AI judgment. Gartner found only 26% of candidates trust AI to evaluate them fairly. Pew Research found 71% of Americans oppose AI making the final hiring decision.
So keep the chatbot in the helper role. Candidates are happy to get an instant answer about shift times or parking. They are not happy to be rejected by one.
What can go wrong with a recruiting chatbot?
Security. In 2025, security researchers found a test account on McHire, the McDonald’s hiring site run on Paradox, protected by the password “123456”. Combined with a second flaw, it could have exposed up to 64 million applicant records, according to CSO Online. Paradox fixed it within a day and said only the researchers viewed any data. The lesson for buyers: a chatbot holds a lot of personal data, so ask vendors about security testing, as covered in Module 5, Lesson 2.
Dead ends. A bot that cannot understand a question and keeps repeating the menu is worse than no bot.
Wrong answers. A generative bot that is not limited to your own content may invent a benefit or a pay rate.
Accessibility. The Justice Department’s ADA guidance says employers must offer reasonable accommodations and remain responsible even when they use another company’s hiring technology. Candidates must be able to reach a person another way.
What rules apply to chatbots and texting?
| Topic | What to know |
|---|---|
| Saying it is a bot | California’s bot disclosure law targets bots used to sell or sway votes, not hiring. Disclose anyway; it is simply good practice and builds trust |
| Texting consent | Since April 2025, US rules let people opt out of texts by any reasonable means, such as replying STOP or QUIT, and you must honor it within 10 business days |
| AI hiring laws | If the bot screens or scores candidates, state and city AI hiring laws may apply. See Module 4, Lesson 3 |
| Telling candidates | Explain where AI is used and how to reach a person. Module 4, Lesson 4 has a sample notice |
What should a chatbot hand to a human?
| Let the bot handle | Hand to a person |
|---|---|
| Questions about the role, location, shifts and process | Pay negotiation and offer questions |
| Application status updates | Requests for accommodation |
| Booking, reminders and reschedules | Complaints and anything a candidate says is urgent |
| Basic knockout questions, such as work authorization or shift availability | Every rejection decision |
How do you set one up well?
Start with your top 20 questions. Pull them from recruiter inboxes and write clear answers. That is the bot’s knowledge base.
Limit it to your content. Tell the vendor you want answers drawn only from your approved material, and a fallback of “let me connect you with a recruiter”.
Test it as a candidate. Ask awkward questions, misspell things and ask about pay. Fix every dead end before launch.
Read the transcripts weekly. Unanswered questions show you what to add and where the process confuses people.
Quick answers
Do I need a chatbot if I hire a few people a year?
Probably not. A clear careers page, a self booking link and fast personal replies do the job. Chatbots pay off at volume.
Can a chatbot reject candidates?
Some can ask knockout questions, such as whether someone can work weekends. Keep final rejections with a person, and make sure knockout questions are truly required for the job.
Is texting candidates better than email?
For hourly and frontline roles, usually, because many candidates apply on their phones. Get consent first and honor opt outs quickly.
Should the bot have a name and personality?
A friendly name is fine, as long as the bot says clearly that it is automated.
Explore in the directory
Module 6: AI across the recruiting desk
- How can AI help with hiring manager intake meetings?
- How do you write job descriptions and job ads with AI?
- What is a recruiting chatbot, and does it help candidates?
- How does AI help with reference checks and job offers?
- How do agency recruiters use AI to win new clients?