Advanced keyword mapping is similar to regular keyword mapping, but instead of users manually mapping keywords to related terms, our NLP tool automates the process.  During a Job Search conversation, your client's AI Assistant will ask candidates what jobs they are interested in at the client’s company and/or for a specific location. When the candidate responds with a job title, this will initiate the advanced keyword mapping process.

Advanced keyword mapping logic

1. During a Job Search conversation, your AI Assistant will ask candidates what jobs they are interested in at your company and/or for a specific location.
2. The candidate will respond with a job title.
3. This job title sent will then be searched against alternate titles in the O*NET Occupation dataset, which includes a wide range of job titles and descriptions from various industries.
4. This system will identify and match this job title keyword with other titles that are similar or closely related.
5. In real-time, the NLP will choose the most relevant keywords associated with that job title.
6. The NLP will then have your AI Assistant suggest those job titles to the candidate for them to consider for the actual job search, helping them find relevant job postings even if the exact title they searched for is not included in the feed.

Example: If a candidate is looking for “Baker” jobs, your AI Assistant can suggest other job titles like "Pastry Chef" or "Bakery Associate" to consider for the actual job search. 

Location context

NLP will use location context within the conversation to directly connect location names to physical places using Geo Search to identify location names and postal codes more effectively, eliminating confusion when recognizing different places in our NLP system. This includes the ability to comprehend state abbreviations (i.e. “MA jobs” as “jobs in Massachusetts,” not “medical assistant positions”) and determine city names that are similar to job titles (i.e., “Baker, CA” vs. “Pastry Cook”).

If no location context is provided, your AI Assistant will send a “Job Context” prompt (“Sounds good. What are you looking to do at [company name]?”).

Job context

NLP will use job context within the conversation, where the NLP analyzes a candidate's answer to “What are you looking to do at the company?”

In this context, the focus is on understanding the candidate’s desired role within the specified company using Brand Name, O*NET, Job Entity extractors, giving prominence to filters like “Part-Time” or “Student” for search refinement while making location completely irrelevant.

No context

When context is not available to help with the decision-making process, the NLP should extract all relevant information from the conversation and continue the conversation by prompting the candidate to choose from available options that can clarify the intent.

While your AI Assistant has a prompt for candidates to select an exact location when multiple options are found, they don’t yet have a similar prompt for other types of entities. 

• Example: In cases where different entities partially overlap in the text (Job Location v.s. Job Category), we prioritize the longer value (e.g., “Baker City” would resolve to “Job Location” rather than “Job Category”).

Inferred context

At times, the NLP may be able to use previous information shared in the conversation to help provide the candidate with a relevant keyword mapping. In pulling together these pieces of information, this will decrease the chances of ambiguity and allow the NLP to infer the context around the job search. One particular example is the frequent confusion between location and brand entities, which can be resolved using this logic.