7 best practices when using recruitment chatbots in the age of AI

Chatbots can help recruiters with tasks like scheduling interviews, but the CHRO should require their team to take precautions when rolling out a recruitment chatbot.

Chatbots could send incorrect replies to candidates or make larger mistakes, such as sending a job offer to the wrong candidate or sharing PII data. Providing an incorrect response could lead to an unhappy candidate, while serious errors could lead to lawsuits or fines.

Here are some best practices for using recruitment chatbots that CHROs should ask their team to follow.

1.    Carefully evaluate the chatbot vendor

Before selecting a chatbot, a company must make sure that the product is the right fit for their particular organization.

CHROs and their subordinates should review the vendor’s documentation, reach out to the vendor’s references, confirm that the chatbot can integrate with the company’s existing ATS as well as other systems and test the chatbot.

2.    Be transparent about company AI use

CHROs must make sure that their subordinates understand how the chatbot is gathering data, how the AI is analyzing data and how the AI determines what it will share with the requester so employees can confirm the chatbot’s factual accuracy and adjust if necessary.

For example, if the AI is providing incorrect or insufficient responses, the HR team can provide the chatbot with additional data or have the AI go through more training.

Candidates should also be aware when they are chatting with a bot instead of a real person.

3.    Prevent potential privacy issues

CHROs must ensure the chatbot is not sharing confidential information. For example, the chatbot should not be able to send the questions that will be asked during the interview or the available compensation for a certain role to a candidate.

Also, managers should not be able to view confidential information about candidates and employees when using the recruitment chatbot.

4.    Take action against AI bias in recruiting

AI bias in recruiting is an ongoing issue, and CHROs must ensure their company is preventing it from happening to the extent possible.

CHROs should ask their subordinates to continually confirm that the chatbot is not using demographic data to discriminate against candidates. In addition, the HR team must feel confident that pre-screening notes and candidate rankings are free of bias.

5.    Create a clear escalation path for the chatbot

The chatbot must be trained on situations that must be passed on to a human.

For example, a chatbot must send the situation to a human employee when it involves making an offer to a candidate, recommending compensation targets or sharing information about the recruiter assigned to the job requisition.

6.    Test the chatbot

A chatbot requires a certain amount of testing to be successful. CHROs should ensure the chatbot has been tested from the point of view of candidates, managers and existing employees.

Once the chatbot goes live, employees must carry out regular spot checks to confirm that the chatbot is referencing the right information and only responding with appropriate answers.

7.    Properly train recruiters

The recruitment team must be trained so they can maximize the effectiveness of the chatbot. Recruiters must understand the tasks that the chatbot can perform as well as its limitations.

Training can include sharing and explaining some good templates for corresponding with the chatbot as well as discussing how recruiters can help improve the chatbot’s replies to candidates.

Eric St-Jean is an independent consultant with a particular focus on HR technology, project management and Microsoft Excel training and automation. He writes about numerous business and technology areas.