There are two pieces to Q&A, which work together to provide the proper response for candidates:

1. Automated NLP lookup process: When a candidate engages with your AI Assistant, we send their phrase through our NLP apps to determine the best match to provide the appropriate answer.
   • This process becomes even more complex as you add custom contents, group targeting, location targeting, geo-targeted content, and multilingual capabilities to provide more targeted messaging.
2. Content Management: Once your AI Assistant matches the appropriate topic, they will pull the approved response from the Candidate Q&A database that is in place for the client to send to the candidate.

To begin the implementation process for Job Search, we will either connect or scrape their job feed so that we have access to all of their jobs in our database. Once this feed is set up, we can begin setting up Job Search in the CEM.

There are two points where your AI Assistant uses NLP to provide Job Search results:

1. Your AI Assistant uses Candidate Q&A logic to first understand that the candidate would like assistance with their job search.
2. Your AI Assistant then uses NLP and Job Search logic to determine which job results to provide to the candidate.

Your AI Assistant can understand and communicate with candidates to answer thousands of questions each day, in over 36 different languages. The system utilizes both language detection technology and NLP to provide candidates with a positive experience in their language of preference.

• Our technology can detect a candidate’s preferred language from the initial interaction.
• Your AI Assistant can also switch languages mid-conversation by understanding a candidate’s language request using NLP.

Our Screening and Scheduling products are supported by both NLP and other technology built to support both of these processes.

• Screening: Yes/No criteria in response to screening questions.
• Scheduling: Selecting interview times and reschedule/cancel requests.

Your AI Assistant has been trained on the following in regards to Scheduling:

• They will recognize when a candidate responds saying that they are unavailable to meet for a specific interview time after your AI Assistant has provided interview times.
• They were trained on understanding a candidate’s preferred time when they include a space between the numbers of the time selection.

Profanity refers to language that is considered offensive or vulgar and includes words or phrases that are socially unacceptable, taboo, or inappropriate in certain contexts or cultures. Currently, the dialogues with your AI Assistant are being monitored for offensive language and profane words through a blacklist of words that are deemed inappropriate or offensive.

However, your AI Assistant is also trained on certain words that may appear offensive but are not in the right context. These include:

• Candidate’s name
• Job title
• Address
• Phone Number
• Email Address

If a candidate sends a message containing a profane word and a job title, then your AI Assistant will not repeat the [profane word + job title] in their message and, rather, complete the intent (e.g. job search, confusion on request, etc.).