An automated hiring pipeline is the series of "if this, then that" steps your team currently does by hand (sending assessments, moving candidates on, offering interview times, chasing people who go quiet) set to run on their own. You design it as triggers, actions and logic, and in the 100Networks Workspace you build it without code and approve exactly what it will do before it touches a real candidate.
The Workspace is 100Networks' AI hiring team for the roles a company fills itself. It is not an applicant tracking system; it works alongside the one you have. This post is the practical part: how to design your first automation, whatever tool you use, and how it works in the Workspace.
Your pipeline is already a workflow, run by hand
Think about what happens for every applicant today. Someone reads the CV, sends a screening interview, waits, checks the result, offers interview times if they pass, sends a decline if they do not, and nudges the ones who go quiet. That is a workflow: a set of triggers and conditional actions. The only problem is that a person is the runtime, doing it over and over, slightly differently each time, and dropping steps in a busy week.
Automating it does not change the logic. It makes the logic run the same way every time.
The three building blocks
Every hiring automation is made of three kinds of step.
- Triggers start it. In the Workspace these include: someone applies, a sourced candidate is added, a candidate's stage changes, an AI interview is finished, an interview is booked or takes place, a candidate replies to outreach, a candidate goes quiet, a tag is added, a background check finishes, an offer is accepted, or you run it by hand.
- Actions do something: send an email, send an AI interview invite, send interview times to book, move a candidate to another stage, decline by email, archive, tag or flag, add a note, add to a talent pool, assign an owner, notify your team, post in Slack, add someone to an outreach sequence or stop one, request references, start a background check, extend an offer, or send data to another system.
- Logic decides or waits: branch on a condition, switch three ways on a score, filter, take a random sample, wait for an event, wait a set time, or end.
Events come in, logic routes them, actions fire.
Build one: a first-round screening loop
Here is a first automation worth building, step by step:
- Trigger: someone applies to the role.
- Action: send an AI interview invite. (For how the interview itself works, see how AI interviews work.)
- Wait for event: the AI interview is finished, with a timeout of, say, three days.
- Switch on score:
- High: move the candidate to the next stage and send interview times to book.
- Middle: tag them for a person to review, and notify the team.
- Low: send a respectful decline.
Every applicant gets the same prompt response and the same criteria, and nobody falls through the cracks because a recruiter was in back-to-back meetings. The middle band is deliberate: a score orders a queue, it should not make every decision.
Approve it before it acts
An automation that emails real people needs a deliberate switch-on, not a save button. In the Workspace:
- It starts as a draft. A draft never acts on real people.
- You see what it would have done. Before it goes live, it is previewed on the last ten candidates its trigger would have seen. Nothing is executed.
- You choose its scope. It starts from "only the roles I choose", never every role.
- It says what it will do. The switch-on spells out only what this automation's steps can do, for example "send emails and decline people by email".
Making it live is the team's approval. This is the Workspace's general rule applied to automations: nothing reaches a candidate until your team approves it. You can also run a live automation once on one candidate you pick, which is the best way to test a change.
Change it in plain words
You do not have to rebuild a workflow to change it. Describe the change, such as "wait two days before the reminder", "only for senior roles" or "also post it to Slack", and the Workspace proposes the new steps first. Nothing changes until you apply it, and if the automation is live, the screen says so before you do.
It is durable, not fire-and-forget
A real hiring workflow runs for days or weeks per candidate, so it cannot live in memory. A run that is waiting for an interview, a reply or a set time saves its place and picks up where it left off, and a step that fails is retried. Each automation shows its runs, how many finished and how many failed, and where people go in and come out at each step, so you can see where your funnel actually leaks.
Automation handles the plumbing; people keep the judgment
The goal is not to remove people from hiring. It is to remove people from the repetitive plumbing: sending, waiting, nudging, moving candidates on clear criteria. The judgment stays with you. For anything else, Pilot, the Workspace's plain-words assistant, proposes the action for a person to confirm.
Once your first automation is running you will spot five more worth building. Start with the one that wastes the most of your week. For the measures worth watching once it is live, see recruiting analytics that matter.
And when a role is too hard for any pipeline, because the right person is not applying at all, that is what the 100Networks marketplace is for: send us the role, specialist recruiters who have filled it before work it, and you pay 8.33% of first-year salary only when the person joins. See how it works for companies.