Aipedia: Recruitment agencies

Recruitment chatbots and automated outreach

Aipedia, by the Apex broking team · Last reviewed 2026-10-27

An agency is responsible for what its recruitment chatbot tells candidates, including what it says about roles, pay and next steps. Bots can screen and message thousands of people, so one bad script can produce many complaints at once.

Part of: AI in recruitment agencies

In short

Agencies may use AI chatbots to answer candidate questions, run screening questions, book interviews and send automated outreach by email, text and messaging apps. A generative bot can state pay, hours or contract terms that the client never agreed, or tell a candidate they are rejected without the safeguards UK GDPR now requires. Outreach that promotes the agency's services may engage PECR regulation 22 consent rules. Since 19 June 2026, data protection complaints must be acknowledged within 30 days. Whether PI responds to a chatbot complaint depends on whether there is a professional service loss and on the wording.

How agencies use chatbots

Recruitment chatbots sit on careers pages, job boards and messaging apps. They answer questions about vacancies, ask knock-out screening questions such as availability or licences, collect CVs, schedule interviews and send status updates. Outreach tools draft and send personalised messages to sourced candidates at scale. Generative chatbots answer in free text, which makes them more natural and less predictable.

Some agencies also run chatbots under a client's brand as part of a recruitment process outsourcing service. The candidate may then think they are dealing with the employer, while the agency configures and runs the bot. That makes it important to agree in writing who approves the scripts, who answers complaints and who is controller for the conversation data.

The duties that apply

Accuracy and honesty: REC Code principle 2 requires honesty and transparency. Under CAP Code section 20, job ads must describe conditions accurately and state earnings precisely; a bot answering questions about a vacancy should be held to the same standard.
Automated decisions: a bot that rejects candidates on screening answers without meaningful human involvement is making a significant decision, so Article 22C safeguards apply: information, representations, human intervention and a way to contest.
PECR: under regulation 22, marketing emails and texts to individuals generally need specific consent; the soft opt-in does not cover prospective customers or bought-in lists, and you must not hide your identity and must give an opt-out. Whether a particular job message is direct marketing depends on what it promotes.
Data protection: privacy information for data the bot collects, reg 28 confidentiality, and since 19 June 2026 a complaints procedure with acknowledgement within 30 days.

What can go wrong

Illustrative: an agency's chatbot, asked about a warehouse role, tells candidates the job pays overtime at a higher rate and includes guaranteed hours. Neither is true. Several candidates resign from other jobs to start. When they find out, they complain to the agency and to the Fair Work Agency, and post screenshots online. Separately, the bot's knock-out questions reject anyone without a driving licence for a role where the client said a licence was preferred, not required. The client loses good applicants and blames the agency.

Complaints that end with the agency

Candidates rarely distinguish between the agency and its software supplier. Complaints arrive at the agency, regulators look at the agency, and clients hold the agency to its service. Supplier contracts often limit what the agency can recover. The ICO's Recruitment rewired report said it will scrutinise recruitment platforms as well as employers, so the supplier may face questions too.

How PI usually responds

There is no statutory PI requirement for agencies. A client's claim that a chatbot misdescribed or wrongly screened candidates, causing a failed placement campaign, may fall within a recruiter's PI, subject to the wording. A candidate's claim about a misleading promise is less clear: it depends on whether a duty was owed and whether the policy covers claims by work-seekers. PECR and data protection fines are commonly excluded. If you build or sell chatbots, consider specific chatbot insurance.

Practical controls

Limit the bot to approved vacancy information; do not let it generate pay or terms.
Tell users they are talking to an automated system and how to reach a person.
Check knock-out questions against the client's written requirements.
Give rejected candidates a route to human review.
Check consent and opt-out handling for automated outreach.
Keep conversation logs and review a sample weekly.

How Apex helps

Apex arranges PI for recruitment agencies; a broker can check how your PI and cyber policies would respond to chatbot complaints. See the AI in recruitment agencies hub, telling candidates you use AI, AI CV screening and AI-written job adverts.

Sources

Frequently asked

Is an agency liable for what its chatbot says?

In practice, yes. The chatbot speaks for the agency, and REC Code principle 2 requires honesty and transparency. A false statement about pay or terms can lead to complaints, client disputes and possibly claims, whoever supplied the software.

Do PECR rules apply to recruitment outreach?

They may. PECR regulation 22 requires specific consent for marketing emails and texts to individuals, with a limited soft opt-in that does not cover prospective customers or bought-in lists. Whether a message is marketing depends on what it promotes, so check each campaign.

Can a chatbot reject candidates automatically?

It can screen, but an automated rejection without meaningful human involvement is a significant decision under UK GDPR. The candidate must get Article 22C safeguards, including human intervention and a way to contest. Build in a route to a person.

How quickly must we respond to a data protection complaint about our bot?

Since 19 June 2026, organisations must have a data protection complaints procedure and acknowledge complaints within 30 days. Make sure chatbot complaints reach that procedure rather than staying in the bot.

Check your PI covers how you use AI

Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.

Start a PI quote Request a callback

Apex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.