AI and Candidate Experience: Chatbots, Updates and Transparency

By Kevin Kim · Updated Oct 11, 2026

AI and candidate experience is the question of how automation in hiring (chatbots that answer applicant questions, systems that send status updates, tools that schedule interviews and screen applications) changes what it feels like to apply for a job. Done well, AI gives applicants fast answers and a clear view of where they stand. Done badly, it puts a machine between the candidate and any human who can explain a decision, and in several jurisdictions it also triggers legal duties to tell people the machine is there.

This guide is for HR directors, talent acquisition leaders and recruiting managers who already use, or are about to use, AI somewhere between "apply" and "offer". It's part of the recruiting section of the AI for HR collection, so start with AI in recruiting for the whole picture. Here the focus is the applicant's side of the screen. Nothing below is legal advice; rules differ by place and change, so check with counsel before you rely on any of it.

Key Facts

• In a Pew Research Center survey of 11,004 U.S. adults fielded December 12 to 18, 2022, 66% said they would not want to apply for a job with an employer that uses AI to help make hiring decisions (Pew Research Center, April 2023).
• The same Pew survey found Americans opposed AI making a final hiring decision by 71% to 7%, yet 47% thought AI would do better than humans at treating all applicants the same way, against 15% who thought it would do worse (Pew Research Center).
• New York City requires employers to give candidates notice 10 business days before using an automated employment decision tool, and the city began enforcing the rule on July 5, 2023 (NYC Department of Consumer and Worker Protection).
• Under the EU AI Act, chatbots must be designed so people know they're talking to an AI unless that's obvious, and the transparency rules apply from 2 August 2026 (EU AI Act, Article 50). Rules for high-risk employment systems apply from 2 December 2027 (European Commission).

What Candidate Experience Means Once AI Is Involved

Candidate experience is the sum of every contact an applicant has with your hiring process: finding the role, applying, hearing back, interviewing, getting a decision. Before AI, the common complaints were silence and delay. Applicants sent a resume and heard nothing for three weeks, then got a form rejection.

AI changes three parts of that journey.

• Answering questions. A retrieval-based assistant answers from your own careers page, benefits summary and process documents ("Is this role hybrid?", "What happens after the phone screen?"). In ACE Framework terms this is the rag-assistant pattern: it fetches from approved sources instead of improvising.
• Keeping people informed. Workflow automation sends a confirmation on application, moves a candidate to the next stage and sends the update, and nudges a recruiter when someone has waited too long. This is the workflow-copilot pattern: software does the routing and drafting, a person owns the exceptions.
• Moving faster. Scheduling, reminders and first-pass sorting shorten the gaps between steps, which is usually what candidates feel most.

What AI doesn't change is the thing candidates care about most: whether a real person looked at them fairly and can explain what happened. That's why the research below keeps returning to humans and honesty rather than speed.

What Candidates Think About AI in Hiring

The best-known independent measure is the Pew Research Center's survey of U.S. adults. The fieldwork was in late 2022, so treat it as a baseline rather than a current reading.

The results are mixed in a useful way. Two-thirds (66%) said they wouldn't want to apply to an employer using AI in hiring decisions, and the public opposed AI making the final hiring call by 71% to 7% (Pew Research Center). But 47% also thought AI would do better than people at treating every applicant the same, versus 15% who thought it would do worse. So candidates aren't simply anti-machine. Many believe a machine could be fairer than a hurried recruiter, and they still don't want a machine to be the one who decides.

That fits how most people behave in practice. A chatbot that books an interview feels like a convenience. A system that quietly ranks you and decides you're out feels like a verdict from nobody. The practical reading for a manager: use AI generously for logistics and answers, use it carefully for evaluation, and keep a human visibly responsible for the outcome.

Be wary of vendor-run surveys that report very high comfort with AI in hiring. Many are small, and the question wording often asks about convenience tasks, which almost everyone accepts. A fair test of any number is who asked, how many they asked and what exactly they asked.

Real Examples by Function

Talent acquisition. A recruiter handling 40 open roles gets the same ten questions from every applicant. An assistant trained on the careers site and the interview process answers them at 11 p.m., links the source, and hands off to a recruiter when the question is about a specific application, an accommodation or a complaint.

Recruiting operations. Scheduling is the cleanest win. The tool offers slots from the interviewers' calendars, confirms by message, reminds the candidate the day before and reschedules without three rounds of email.

Hiring managers. A manager who owes feedback on four finalists gets an automatic prompt after 48 hours, and the candidate gets a status message saying the decision is pending. The AI isn't deciding anything; it's removing the silence.

Employer branding and communications. Generative tools draft rejection messages that name the stage reached and the next step for people who might apply again. A recruiter reviews and sends, so the message isn't a template pasted to everyone.

Compliance and legal. For tools that score or rank applicants, compliance owns the notice text, the audit record and the alternative path. They also own the question of whether an assistant's wording could be read as a promise or a decision.

For a deeper look at the screening side, see AI resume screening, and for the agent version of a first-contact screener, see the AI recruiting screener agent.

Where Candidate Experience Goes Wrong

The Disclosure Rules You Need to Know

Several places now require employers to say when AI is part of hiring. The details differ, so treat this as a map and not a compliance plan.

New York City, Local Law 144. If you use an automated employment decision tool to screen candidates for jobs in the city, the tool must have had a bias audit within a year of use, a summary of the results must be public, and candidates must get notice. The city says the notice must come 10 business days before the tool is used, and it began enforcing the law on July 5, 2023 (NYC Department of Consumer and Worker Protection).

Illinois, Artificial Intelligence Video Interview Act. For Illinois-based positions where applicants record video interviews and AI analyzes them, the employer must notify each applicant beforehand that AI may be used to consider their fitness for the position, explain how the AI works and what general characteristics it evaluates, and obtain the applicant's consent. Employers may not use AI to evaluate applicants who haven't consented (820 ILCS 42/5 via FindLaw).

EU AI Act, Article 50. Providers must design AI systems that interact directly with people so that those people are told they're dealing with an AI, unless that's obvious to a reasonably well-informed person, with application from 2 August 2026 (EU AI Act, Article 50). In practice, a recruiting chatbot on your careers site should introduce itself as automated.

EU AI Act, Article 26. For high-risk systems, which include AI tools for employment and worker management such as CV-sorting software, employers who deploy them must inform workers' representatives and affected workers before use, and must inform people who are subject to decisions made or assisted by such a system (EU AI Act, Article 26). The Commission says these high-risk rules apply from 2 December 2027, after the AI Omnibus extension (European Commission).

Several US states have passed or proposed more rules on AI in employment, and some have moved their dates. If you hire across states or countries, keep a short, current register of what applies where and who owns each item. The principle underneath all of it is the same: tell people early, in plain language, before the step happens.

How to Improve Candidate Experience With AI, Step by Step

1. Map the journey. List every touchpoint from job post to offer and mark where candidates currently wait or ask questions. Those gaps are your candidates for automation.
2. Sort touchpoints by risk. Questions, scheduling and status messages are low risk. Anything that ranks, scores or filters people is high risk and needs more care before you automate it.
3. Ground the assistant in approved content. Feed it your careers page, process description and benefits summary only. If the answer isn't in the sources, it should say so and hand off.
4. Write the disclosure first. Draft the plain-language notice ("This assistant is automated. Recruiters make every hiring decision.") before building anything, and place it where the interaction starts.
5. Build the human exit. Every automated step needs a way to reach a person, a named owner and a response time you can keep.
6. Add human review where it matters. For any step that influences who advances, set a real checkpoint, as described in human-in-the-loop. Be able to explain the factors behind a recommendation, the idea behind explainable AI.
7. Test before launch. Run realistic applications through the process, include candidates with accents, gaps in work history and non-standard resumes, and compare outcomes across groups.
8. Pilot one role family. Measure and listen for a full hiring cycle before expanding.
9. Review on a schedule. Check complaint logs, abandoned conversations, correction rates and fairness results quarterly.

What to Measure

Track speed and experience together, because a fast process that candidates dislike hurts your employer brand.

• Responsiveness: time to first reply, time between stages, share of applicants who receive a status update.
• Human access: how often candidates ask for a person, and how long it takes to reach one.
• Quality: share of chatbot answers a recruiter had to correct, and the number of conversations that ended without an answer.
• Fairness: stage-by-stage pass rates by group for any screening or ranking step.
• Sentiment: a one-question survey after a decision, sent to hired and rejected candidates alike.

If a metric improves and complaints rise, believe the complaints.

Common Pitfalls

• Treating transparency as a footer link. A notice buried in terms and conditions doesn't count as telling candidates. Put it in the step where AI is used.
• Letting speed substitute for substance. An instant rejection with no explanation feels worse than a slow one.
• Copying one region's rules everywhere. A notice written for New York may not satisfy Illinois consent or EU duties.
• Forgetting accessibility. Candidates with disabilities may need an alternative to a chatbot, a timed assessment or a recorded interview. Make the request path obvious.
• No ownership. If nobody owns the assistant's content and the notice text, both go stale. Name an owner and review them on a schedule.
• Reading the Pew result as permission. That 47% think AI could be more even-handed doesn't mean they'll accept an unexplained machine verdict. Keep the human accountable and visible.

For the wider decision about which hiring tasks to automate at all, read AI across the employee lifecycle, and for help comparing tools see how to choose recruiting software.