Purpose

This article explains how to use AI-Powered Calibration to find the best-matched candidates from your internal talent pool. You'll learn how each calibration setting works, how it influences the AI's Match Score, and how to interpret the results to fill your roles faster and more effectively.

What is Calibration?

Calibration is the process of defining and refining role requirements. This process, led by the recruiter with input from the hiring manager, helps guide the Eightfold AI to match candidates based on current, relevant criteria.

By defining the criteria most important to the hiring leader, calibration becomes the foundation of the AI's power within Eightfold.

Introduction

Large organizations have massive databases of past applicants and former employees. Our AI surfaces the most relevant talent from this existing pool instantly.

When you create a position, the Leads tab is automatically populated with a stack-ranked list of matched candidates. The key to ensuring these leads are high-quality lies in Calibration - the process of teaching the AI exactly what you're looking for.

Understanding Calibration

Calibration captures job requirements. It guides the AI to ensure candidates in your Leads pipeline are the best fit.

You can access this page by clicking the Calibrate Position button on any position's overview page.

Calibrating a Position

Each field on the Calibration page serves a specific purpose, either by acting as a hard filter that narrows your pool or by influencing the Match Score calculation.

Who do you want to Hire?

This section provides the most basic and crucial information about the role.

1. Position

This is the most important field as it's the primary key for the AI to search and match candidates. While it defaults from the ATS, you should always review it for accuracy.

2. Job Location

Enter the exact work location at the city level. To ensure accuracy, select the option from the dropdown that includes the state or country (e.g., "San Francisco, CA"). The location field impacts the match score by restricting the candidate pool.

This field acts as a hard filter that determines the geographical scope of your search. It doesn't affect the Match Score of candidates who pass the filter, but it'll exclude those who don't.

• Functionality: Enter one or more cities for the position.
  
• How it Works:
  
  • For Onsite or Hybrid roles, the system fetches candidates within a 100km radius of the specified location(s).
    
  • You can broaden the search using the Workplace Type setting under "Advanced Options." Selecting Relocation Allowed or Remote within Country searches the entire country, while Remote Globally removes all location restrictions.
    
• Best Practices:
  
  • Multiple locations are fine.
    
  • For U.S. roles, always include the state.
    
  • Avoid using specific street addresses.

Who is your ideal candidate?

Adding ideal candidates is the most impactful step in the calibration process. The AI uses these profiles as a learning tool to understand what a good fit looks like.

1. Ideal Candidates

Ideal candidates have the most significant impact on the match score. Their skills and experience establish the foundation for the rest of the calibration. Strong ideal candidates with diverse yet complementary backgrounds in skills and experience are a plus.

• Functionality: Add 3-5 profiles of individuals who represent the perfect fit for the role. These can be strong current employees, exemplary past applicants, or sourced leads.
  
• How it Works: The AI analyzes the skills, experience, education, and career trajectories of these ideal candidates to create a highly accurate benchmark. It then finds other candidates in your talent pool with similar profiles.
  
• Selection Best Practices:
  
  • Choose strong performers: Select individuals who are currently in the same or a similar position and have the skills and experience you're looking for.
    
  • Rich profiles: Ensure the candidates have rich, detailed profiles that provide the system with good data.
    
  • Quantity: Select at least 3, but no more than 5. Having more than five can make it difficult for the AI to focus on specific preferences.
    
  • Avoid poor fits: Don't select individuals who are no longer a good fit for the role. This can negatively impact match scores

When selecting Ideal Candidates, the goal is to define the profile of a strong performer for the role. Follow these best practices to ensure a high-quality calibration:

• Quantity: Select a minimum of three, but no more than five, ideal candidates. Choosing more than five can make it difficult to focus on specific requirements and preferences.

• Source:

  • Start by identifying employees currently in the role at your company who have rich profiles detailing their skills and experience.
  • If you don't have enough internal candidates, look for individuals on public platforms like LinkedIn who are in similar roles, have the required skills, and have robust profiles.
  • For new roles that have not previously existed in the organization, it's recommended to look entirely outside the company for ideal candidates.

• Broad Representation:

  • If a role spans multiple business sectors (e.g., Academy Consultants with backgrounds in Business, Healthcare, or IT), choose an ideal candidate from each area. All candidates will be matched against the same combination of all ideal candidates, ensuring a broad and inclusive match.
  • If you are looking for specific skills, such as Ruby, include an ideal candidate who has that skill in their profile.

• What to Avoid:

  • Don't select individuals who are not relevant to the role. Doing so may introduce irrelevant or inflated skills and experiences that negatively impact the match scores.
  • While you should not select individuals who are no longer performing in the role, you should consider the skills of those who have moved on to other roles to find strong candidates. This is particularly useful in roles where career progression is important.

• When to Change Ideal Candidates:

  • If a position has been open for a long time, check the ideal candidate profiles. Ongoing profile refreshes may have changed their skills or experience, which could impact the calibration.
    
  • When you copy a calibration from another position, always review the selected ideal candidates. Ensure they are relevant to the new position's specific requirements.

Refine Requirements

This section allows you to provide more specific details that complement your ideal candidate selections. The selection of Skills and Experience criteria should complement the Ideal Candidate group. Avoid adding skills and experiences that are unrelated or contradict those of the ideal candidates, as this can negatively impact the candidate pipeline and match scores.

1. Skills/Keywords

Define the core competencies required for the job. This section has a high impact on both fetching candidates and calculating the Match Score.

• Functionality: Add essential skills for the role. Our AI uses a multidimensional model to understand related skills and context, not just keyword matches (e.g., it knows that a candidate with "C++" and "Machine Learning" experience can likely learn "Python").
  
• Settings:
  
  • Preferred (Default): The skill is desirable and will positively influence a candidate's score. This is the best option for most cases.
    
  • Required: A hard filter for must-have skills. Only candidates with this exact skill will appear in the Leads pipeline. Use this sparingly, as it can severely limit your results.
    
  • Exclude: A hard filter that removes any candidate who possesses this skill.

Note: Required vs. Preferred: Skills are defaulted to Preferred, which is the most inclusive and recommended setting. The AI will look for a semantic match, not just an exact word match. Use Required sparingly, only for skills that are absolute must-haves. When a skill is required, the AI looks for an exact match and won't consider inferred skills, which can significantly narrow your talent pool.

Preferred Skills

• To ensure accurate calibration, select a minimum of three hard skills for the ideal candidate. Avoid using broad skills that may skew results.
  
• Review the related skills for each selection to ensure relevancy and to identify potential outliers that could impact calibration and matching.

Required Skills

• Limit required skills to one if possible. A skill designated as Required is considered a must-have for the role.
  
• Required skills will disqualify candidates who do not possess them.
  
• The absence of a required skill will also negatively impact an applicant's match score.

Excluding Skills

• Excluding a skill works similarly to a required skill. Candidates with the excluded skill will be disqualified, and their match score will be negatively impacted.

Prevalence: Prevalence provides insight into a skill's availability in the market over time.

Related Skills: Adding related skills gives the system more information for calibration, which leads to more relevant and accurate results.

Top Companies: Reviewing top companies can help you identify where your target talent pool may be found.

2. Experience

This section allows you to define the professional background you are looking for.

• Job Titles: Add at least three relevant job titles. The AI will find candidates with these titles and also identify those with comparable responsibilities, even if the title isn't an exact match.
  Avoid generic titles and acronyms. Like skills, it's not recommended to mark titles as Required, as titling can vary significantly between companies.
  
• Companies: Prioritize candidates who have worked at specific companies (e.g., competitors).
  Note: You can mark specific companies or industries as Preferred to give candidates a slight boost, or as Required to narrow the pool.
  
• Industries: Narrow the search to candidates with experience in a particular industry.
  

• Preferred Seniority Level: Filter candidates by career level (e.g., Manager, Director, CXO). This is best suited for executive-level or intern roles; for most roles, it's better to leave this open to cast a wider net.
  It's recommended to not use this setting initially. If, after reviewing leads, you find candidates are consistently too senior or too junior, you can apply this filter to better align the talent pool.

  • CXO (C-Suite Professional): A top-level professional.
    
  • VP (Vice President): Holds a high-level management role.
    
  • Director: A professional with managers as direct reports.
    
  • Manager: A professional who has direct reports at a lower level.
    
  • Senior: An experienced individual contributor. This level includes lead roles.
    
  • Mid-Level: An individual contributor with some work experience, but less than a senior member.
    
  • Entry: A recent graduate or someone with 1–2 years of experience.
    
  • Intern: An active student seeking professional experience.

• Years of Experience: Use this to set a range for either "Total Years" or "Years Relevant to this Position". Having more years than preferred has a lesser negative impact on the match score than having too few.

  • Relevant Years of Experience is typically used for specialized, technical, or management roles that require specific training, skills, or certifications.
    
  • Total Years of Experience is more common for non-specialized, non-technical, and individual contributor roles.

Experience, whether relevant or total, will not exclude any candidates. However, it does influence their Match Score. A candidate's score is impacted more negatively by a lack of experience than it is by having more experience than preferred.

3.Education

This section acts as a hard filter to include or exclude candidates based on their academic background. It does not affect the Match Score.

• Functionality: Filter by Schools, Degrees (e.g., Bachelors, Masters), and Graduation Year.
  
• Recommendations:
  
  • For interns, set the degree to Bachelor’s and the graduation year to the current or upcoming year.
    
  • For early career roles, use a Bachelor’s degree with a graduation window from the last two years.

Filtering by Education

Use the Degree and Graduation Year filters to refine your candidate search.

• Degree: The degree filter limits the talent pool to candidates with a specified level of educational attainment.
  
• Graduation Year: This filter is used to refine results for roles that require a specific graduation date, such as Intern or Early Career positions.

Utilizing Degrees

To ensure accurate results, select the appropriate degree and graduation years for your search.

• Intern Positions: Select "Bachelor's" and set the graduation year to the anticipated graduation date.
  
• Early Career Positions: For roles requiring a bachelor's degree, select "Bachelor's" and set the graduation year to align with the required experience level.

A candidate's graduation year is based on the full calendar year.

• Current Year: Including the current year will show candidates who haven't yet graduated but are expected to.
  
• Previous Year: Setting the graduation year to the end of the previous year will exclude candidates who are set to graduate in the current year.

Note: A candidate with a higher-level degree (e.g., a Master's) won't be penalized in the match score. Always validate a candidate's graduation year against their profile or resume.

4. Advanced Options

These settings provide additional controls to fine-tune your candidate pool.

• Location Preferences: Enabling Relocation Allowed will allow remote candidates who are open to relocating to be considered for the role.
  
• Workplace Type: Defines the location requirements. As mentioned earlier, this works with Job Location to filter your pool.
  
  • Onsite
    
  • Hybrid
    
  • Remote Within CountryA
    
  • Remote Globally
    
• Exclude short tenures: A filter to remove candidates who have a history of short tenure in the most recent 3 positions.
  
  • Do not exclude
    
  • Less than 6 months
    
  • Less than 1 year
    
  • Less than 2 years
    

• Settings: Tag the position as Confidential (restricting visibility to assigned users) or a Priority Position

  • Confidential Position
    
  • Priority Position
    
  • Allow as a reference calibration in Calibration Assistant

• Candidate Position Visibility Settings: Enabling/Disabling Accessible via private link only (not searchable without the link) allows candidates to view and apply to this position through a private URL, even if the position is not publicly accessible.

Job Description

This section contains the official job description that candidates will see. While the structured fields above are primary signals, the natural language in the job description is also used by the AI to understand the role's requirements.

• Job Application Link: The direct URL for candidates to apply.
  
• Job Description Editor: A rich text editor to write or paste the job description.
  
• AI Assistant: Use the "Create a Job Description using similar position JD" feature to have the AI generate a description for you based on past, similar roles, saving you time and effort.

Calibration Notes

This is an internal-facing tool designed to improve collaboration and keep a record of the hiring process.

• Functionality: A rich text editor for recruiters and hiring managers to leave internal notes.
  
• How it Works: These notes are not used by the AI to calculate the Match Score. They are strictly for your team's reference.
  
• Best Practices:
  
  • Use this as an "intake tool" to capture key notes from meetings with hiring managers.
    
  • Document the mission, objectives, and milestones for the role.
    
  • Outline the interview process or record any changes to the calibration requirements over time.