Seneca Careers · employer walkthrough

How a candidate shares work, and what you see.

Screenshots from the Seneca Careers demo. Maya Chen applies for a Business Analyst internship at Northgate Analytics and shares four records. Start with what the employer reads, then see how she chose it.

FictionalEveryone and everything here is made up. Maya Chen, Dana Reyes, Northgate Analytics and the role exist only in the demo, which runs on a laptop with no real data. Careers itself is an invite-only pilot, so there is no public sandbox to click through.

Rolling outThe by-activity layout is being switched on for the pilot. Until it is on for a role, an employer sees the same shared records as a plain list, with the same labels. There is no form to fill in before or after reading in either case.

What the employer reads

Dana Reyes (fictional) opens Maya's application. The work Maya shared is there with it: nothing to fill in first.

  1. 01

    The shared work, by activity

    Four activities Northgate confirmed for the role, and under each one the record Maya chose for it. Every row says where the record came from. Nothing is scored or ranked.

    The employer's view of one application. Under the heading Shared work, by activity are four numbered activities. Beside the first: Candidate's words, from a ChatGPT history (file not verified), AI replies shown beside them, with an Open button. Beside the second and fourth: Typed by the candidate, not observed by Seneca. Beside the third: No example shared. That says nothing about ability.
  2. 02

    A record from an AI conversation

    The AI's reply that came before her message, her own words, and the AI's reply after, each labelled and shown open. A line above says other messages in the conversation are not shown.

    The first activity opened. A heading reads: The candidate's quoted messages and the AI replies next to them. Other messages in this conversation are not shown. Then three numbered turns: AI assistant, pointing out a missing cost; Candidate, recomputing the unit cost and margin; AI assistant, confirming the figures.
  3. 03

    Where it came from, and what it cannot show

    Under the words: the file it was taken from and that the file was not verified, what one record cannot establish, and one fixed question to explore in an interview.

    The lower part of the same record. Origin (Seneca): extracted from a ChatGPT chat history file the candidate uploaded; the file was not verified. Limits (Seneca): one record shows one moment and does not establish a skill, a habit or a trait. Something to explore (Seneca, fixed wording): ask what they checked after the assistant's reply.
  4. 04

    An activity with nothing shared

    Maya linked no record to the third activity. The row says so, and says that it tells you nothing about ability.

    The four activity rows. The third, Review forms and reports and confer with management and users, reads: No example shared. That says nothing about ability.
  5. 05

    A record she typed herself

    Labelled as typed by the candidate and not observed by Seneca, with her own context below it and a prompt to ask for the file or a walk-through.

    The fourth activity opened: Query sales data in SQL and reconcile totals with finance. The record, titled by the candidate Revenue query fixed after a reviewer comment, is labelled Typed by the candidate, not observed by Seneca. Below it, Candidate's context in her words.
  6. 06

    When the AI did the work

    One shared record sits under no activity. She only asked, and the AI wrote the result. The record says so plainly, and the AI's text is shown as the AI's.

    Under Other shared work, not linked to an activity: a record with a boxed line reading The candidate asked the AI for this; the result below was written by the AI. Then her request and the AI assistant's reply.

Maya's own record

Before any employer is involved, the record is private to the candidate.

  1. 07

    Her learning records

    Maya sees her own records. Some she typed in herself; others Seneca took from a ChatGPT history she uploaded. No employer can see any of this.

    Maya's learning record page: a list of records, each with a date, a title and a label saying where it came from.
  2. 08

    One record from her ChatGPT history

    Only her own words are quoted. The label says where they came from and what one record cannot show.

    One record in Maya's own list, taken from her ChatGPT history, with her quoted words and an Origin and Limits note.

Choosing what to share

She applies for the internship, then decides what this one employer may see.

  1. 09

    Choose what to share

    She picks which records this one employer may see, and sets when access ends.

    A dialog titled Choose what to share, for Northgate Analytics (fictional), Business Analyst Intern (fictional). A list of her records with checkboxes, four ticked, and an access expiry date.
  2. 10

    Preview the employer's view

    Before anything is shared she sees exactly what the employer will see, and which parts are frozen when she shares.

    A dialog titled This is what the employer will see. Two lines list what is frozen when she shares (the records, her redactions, titles and context, the AI replies, her links) and what is not (her name, headline and application note).
  3. 11

    Put a record under an activity

    For each activity the employer confirmed, she picks at most one of her records, or leaves it empty. Seneca suggests nothing here.

    A panel headed Link your records to this role's activities. Four activities, each with a menu of her selected records. Three have a record chosen; the third is set to No example.
  4. 12

    Redact a passage

    She hides the client's name in her memo. The employer will see [redacted] in its place.

    The preview with one record's text selected and a Redact selected text button; the client's name now reads [redacted].
  5. 13

    Redact inside an AI reply

    Where ChatGPT wrote the text, she can hide a name in the AI's reply, but she cannot remove the reply or its label.

    The record ChatGPT wrote, with the AI assistant's reply shown and one name in it replaced by [redacted by candidate].
  6. 14

    Shared

    Four records, with this employer only, until the date shown. She can withdraw at any time.

    The candidate's Sharing and privacy page: one share with Northgate Analytics, four records, an end date and a Revoke access button.

What the candidate sees, and withdrawing

The candidate sees access, never opinions, and can end it.

  1. 15

    Opened, and when

    Maya sees that the employer opened her records, and when.

    The candidate's Sharing and privacy page showing the share as Opened by the employer, with the time.
  2. 16

    After she withdraws

    The records disappear from the employer's view, and a plain notice says the candidate withdrew sharing.

    The employer's view of the application after withdrawal: no records, and a notice that the candidate withdrew sharing.

On a phone

The same views work on a small screen, for the candidate and the employer.

Phone screen: This is what the employer will see, with what is frozen when she shares.
The candidate's preview of exactly what the employer will see.
Phone screen: the employer's view, Shared work by activity, with the first activity and the record under it.
The employer goes straight to the shared work.
Phone screen: a conversation record with a labelled AI assistant reply and the candidate's words.
The AI's labelled reply, then her words.

Want to see it live?

Kevin can run this same demo with you on a short call, and set up an invite when you're ready to try it on a real role. Candidates start at the candidate page.