Skip to content
Amara AI

Hiring and recruiting

Amara AI helps you define who you want, write the job post, and read every applicant against the same frozen criteria. Every score comes with the evidence behind it. A named person makes every hiring decision.

What it does

Define the role before you post

Start with a weighted profile of the person you want and one record of what the job offers. Everything later in the search is built from those two records. Scores are advisory. They never move, hide, or reject anyone.

  • Build an ideal candidate profile from weighted dimensions. Dimension weights sum to 100, and each attribute is flagged must-have, preferred, or trainable here.
  • Start from a profile template for the dental role, then adjust the dimensions and weights.
  • Freeze the profile before scoring starts. Changing criteria mid-search creates a new version and never edits existing scores.
  • Record benefits and perks once. Only confirmed items can appear in a post or an offer.
  • Get a listing range and an offer range from reported wages, advertised ranges and your own history, each tagged with its date and freshness. Only job-related factors are used, never salary history or protected traits.
  • Market figures come from government wage data and your own entries. Amara AI never scrapes job boards.

Write and check the job post

Amara AI drafts the post from your job facts and reviews it before you publish. You decide when and where it runs.

  • Amara AI drafts the post from one record of job facts, in the tone you pick.
  • A rewrite pass removes cliches and checks that every job fact survived the edit, with a before and after view.
  • A preflight review runs 13 explainable checks, including role clarity, pay, schedule, readability and dental specificity.
  • Each check returns pass, fix, missing, or human review. Exclusionary wording is flagged for a person to review, and Amara AI does not rule on what is legal.
  • Versions for each channel are generated from the same job facts, with a publication status per channel.
  • Publishing and spend always wait for your approval.

Score applicants with evidence

Every applicant is read against the frozen profile, and every score cites the resume text behind it. Scores are advisory. They never move, hide, or reject anyone.

  • Applications arrive from your careers form, a job board apply connection, your hiring inbox, or a manual paste. Duplicates are matched by email and opening, and the source is recorded.
  • Resumes are parsed into a structured profile: experience, dental workflows, systems and credentials.
  • Each attribute is scored 0 to 100 with the resume text cited. An attribute with no evidence scores 0.
  • An evidence ledger marks every criterion as meets, gap, or unclear, with the snippet, source, confidence and weight.
  • The top 3 gaps become questions to probe at interview.
  • Protected traits and pay history are not considered. Sorting or filtering by score happens only when a person chooses it.

Run structured interviews

Interviews use the same questions for every candidate in a role, tied to the profile and to that candidate's gaps. Scores are advisory. They never move, hide, or reject anyone.

  • Schedule interviews by type, time and interviewer. Invitations wait for approval.
  • A question bank by role ties questions to the profile dimensions and to the candidate's gap list.
  • The interviewer scores each rubric line, and the scorecard gives a weighted total out of 100.
  • Recording stays locked until the candidate's consent is recorded. Transcripts are speaker labelled, and every claim in the summary links to a transcript segment.
  • After the interview, an advisory assessment can adjust the match score up or down. The change is shown as a before and after, with the reason, and it is logged.
  • Work samples are scored per rubric dimension, and a side by side view compares application, interview, work sample and verification.

Pipeline, decisions and offers

Amara AI keeps the record and drafts the paperwork. People make the calls. Scores are advisory. They never move, hide, or reject anyone.

  • Candidates sit on a board with stages you configure. Every move is stored with who made it and when, and only a person moves a candidate.
  • The decision loop keeps four stages separate: interview scorecard, advisory recommendation, human decision, offer package.
  • Advance, hold, reject, offer and hire are recorded by a named person only.
  • Offer packages are built from approved pay guidance, confirmed benefits and your job facts, with a versioned history. An offer above the band needs a job-related justification.
  • Offer letters are generated from the approved package and wait for approval before sending. Each carries a fixed notice that it is not legal advice.
  • Talent discovery draws on three sources only: past applicants, referrals from your team, and people you add by hand. Nothing is pulled from the web.

How each score is calculated

Every number shows its weights and its evidence. Scores are advisory. They never move, hide, or reject anyone.

Match score
Each profile attribute is scored 0 to 100 against cited resume text, no evidence scores 0, each dimension is the weighted average of its attributes, and the overall score is the weighted average of the dimensions, shown as strong, good, fair or weak. Scores are advisory. They never move, hide, or reject anyone.
Evidence ledger and completeness
Each criterion is marked meets, gap, or unclear with its snippet, source, confidence and weight, and the ledger shows an evidence completeness percent. Scores are advisory. They never move, hide, or reject anyone.
Advisory recommendation
A five level recommendation from strong hire to strong no hire, with low, moderate or high confidence, the strongest evidence, counter evidence, open unknowns and what a person should verify, drawn only from the frozen profile and recorded evidence. Scores are advisory. They never move, hide, or reject anyone.
Interview scorecard
Every candidate for a role gets the same questions, the interviewer scores each rubric line, and the lines roll up to a weighted total out of 100. Scores are advisory. They never move, hide, or reject anyone.
Interview advisory adjustment
After the interview, an advisory assessment can adjust the resume match score up, down, or not at all, with a magnitude and a reason, shown as a before and after and logged. Scores are advisory. They never move, hide, or reject anyone.
Work sample score
Each assigned work sample is scored per rubric dimension, out of a maximum for that dimension. Scores are advisory. They never move, hide, or reject anyone.
Job post preflight verdict
13 explainable checks each return pass, fix, missing, or human review. The overall verdict is ready, needs revision, or missing information.

What waits for you

  • Your callPublishing a job post, and any spend on a channel.
  • Your callCandidate replies and outreach. Amara AI drafts them, and nothing is sent until you approve.
  • Your callInterview invitations.
  • Your callOffer letters, before they are sent. An offer cannot be approved until a named person has recorded an offer or hire decision.
  • Your callOutreach to a past applicant found through talent discovery. It waits in Approvals as a draft.

Who works on it

Ready to build a stronger dental team?

Start with the agents you need. Add the rest when you are ready.

By continuing you agree to the Terms of Service and acknowledge the Privacy Policy.