AI-generated candidate summary

Recruitment Marketing Public

Overview

From conversations with our customers, we know that screening resumes and qualifying applicants is one of the most time-consuming phases of the hiring process.

The AI-generated candidate summary feature streamlines this step. It gives you a consistent view of every applicant profile and helps you quickly assess applicant job fit before diving deep into the full resume.

This feature uses AI to generate a summary of a resume. This summary is displayed in the candidate's card within Recruitment Marketing, and the summary can also be generated and viewed directly from the candidate grid.

How to use it

Summaries need to be generated on an individual basis and only if the candidate has at least one resume in their files. In case of a candidate having more than one resume, the most recent version will be used for the summary.

Via the candidate card

To generate the summary, simply go to the candidate's card within Recruitment Marketing. Under the candidate's personal information, you will see a "Generate highlights" button as per the screenshot below.

Generate highlights button.png

Clicking on the "Generate highlights" button will trigger the generation of the summary. This might take a few seconds to complete. Once completed, the "Generate highlights" button will be replaced with an accordion containing the candidate's summary.

Highlights accordion closed.png

Click the accordion to display the full summary as seen below (sample resume used for the purpose of creating this documentation):

Candidate highlights accordion.png

Via the candidate grid

In addition to generating and accessing the candidate summary directly from the candidate card, you can also view existing summaries and generate new summaries directly from the candidate grid.

If you haven't done so already, you will need to update your Column Settings to display the Candidate Summary column, which can be done as follows:

  1. From the Candidate page, click the Columns button.
    Candidate Grid Manage Columns
  2. Expand the Candidate section of the column selector.
  3. Scroll down to the Candidate Summary option and toggle this option to be enabled.
    Candidate Grid Candidate Summary Enabled
  4. Your selection will automatically take effect, and the candidate screen will refresh, with the additional column now displayed.

Once enabled, for any candidate who has a resume available against their profile, you will see the option to either:

  • View Summary: Allows you to view an already generated summary. The date the summary was generated will be shown in the pop-up where the summary is displayed.
  • Summarize: Allows you to generate a summary for a candidate. When clicked, the button will temporarily display information to indicate the summary is generating, and once the generation is complete, the button will change to View Summary - allowing you to click to view the newly generated summary.

Candidate Grid Candidate Summary Buttons

If you need to regenerate an existing summary for a candidate, this should be done directly from the candidate card.

Usage guidelines

  • The candidate summary can be regenerated as many times as needed. For your reference, the generation date will be displayed at the bottom of the summary along with the "Regenerate" button.
  • If you cannot see the "Generate highlights" or "Summarize" button for a candidate, that is because that candidate does not have a resume file in Recruitment Marketing.
  • The summary content will be generated by an AI language model and must be reviewed and verified by the user for accuracy, quality and if appropriate, edited before use. It is possible that the AI model might miss key details from the resume, so hiring decisions should not be made based on this summary alone. 
  • You will be able to turn off this feature for your company should you wish to by accessing the AI & ML company settings.

FAQs

Functionality

How is the summary generated by the AI model?
The AI model takes the candidate’s resume as its input. The prompt instructs the solution to summarise the skills, experience and education of the candidate based on the input, and specifically indicates to exclude taking into account factors such as gender, race or ethnicity. 
Can the system detect resumes that are AI-generated or written with the aid of AI?
No, the system is not equipped to detect this. Detecting if content has been created with the aid of AI is not an exact science and has been proven to be ineffective by various tests. Hence, we are not planning to support this as a feature in the near future.
Can customers personalise the summary to include details that they would like to see in it?
No, the solution uses a standard prompt for all candidates.

Hosting, Security, Bias and Privacy

What is the AI model that is used for resume summarisation?
PageUp is leveraging AI models provided via the AWS Bedrock service and hosted on our AWS infrastructure in each region.
Where is the AI model hosted?
The AI model is hosted on PageUp's Amazon Web Services (AWS) infrastructure in each customer's region.
How is the model trained with customer data?
The AI model is not trained on any customer data. PageUp uses the general capabilities of the models available via AWS Bedrock to generate the candidate summaries.
How does PageUp ensure that the AI model is accurate?
PageUp has chosen a commercially available model that performs well on accuracy tasks. It is important to note that this tool is designed to assist, not replace, human judgment. We encourage users to use the summary as a helpful starting point and cross-reference it with the source material to ensure a complete review occurs at key stages in the hiring process.
What control processes has PageUp put in place to ensure that the AI model is not biased?

PageUp is deeply committed to ensuring that our product features are designed to create fair and equitable outcomes for candidates. By using commercially available AI models, PageUp is able to leverage the ongoing advancements made in preventing bias and associated risks by the creators of those models.

The design of this AI-based feature aims to limit bias in the following ways:

  • The AI model is not trained with any data to distinguish between successful or unsuccessful applicants, nor is the model trained or designed to rank applicants.
  • The model parses each candidate in isolation, without any reference to, or comparison with, other candidates that it might have summarised.
  • The model does not retain any of the data it has previously parsed.
  • The summary generated will not include any gender, race or ethnicity information.
  • The source material used to generate the summary is available to viewed by the user via the candidate card within Recruitment Marketing.

Comments

0 comments

Article is closed for comments.