✈Applications

Aurora Hive Sales looks for vacancies for your interim managers, rates the fit with AI and writes draft applications.

The most important point: nothing is sent without your approval. A daily automatic run refills the review list by itself; your work is reviewing, not searching.

Via “Find & prepare applications”. This starts the search and the rating immediately; a few minutes per run.

Aurora Hive Sales searches the web live for vacancies, rates the fit of each interim manager and writes a finished draft for good matches. The same cost rules apply to the search as in the opportunity search.

Job adverts from Google’s job search arrive somewhat later than the other matches, usually after about a minute. So a review list that still looks thin shortly after the start will usually fill up.

They show the path of an application:

For review → Approved → Sent → Declined.

Each card shows the fit score and the reasoning; via “View draft” you get to the finished email and can adjust the subject, text and recipient.

Does it fit? → “Approve”. Does it not fit? → “Decline”.

Enlarge

It is an assessment with reasoning, not a measurement. Always read the reasoning as well: it reveals whether the model has understood the vacancy correctly, and that is where the mistakes show up, not in the number.

Legally, this is the most sensitive point in the whole program: the use of AI in selecting people for positions falls under the high-risk classification of the European AI Act. This entails obligations — human oversight, traceability, informing the people concerned — which need to be clarified before productive use.

That is why human approval cannot be switched off. This is intentional.

In the “Approved” tab: choose the sender profile, then “Send approved”.

These are real applications to real companies. There is no taking them back afterwards.

If your instance is running in protected sending mode, nothing is delivered — then nothing goes out by mistake either. But do not rely on that; check the notice in the header before you send.

Yes — the follow-up reminds you by itself. You tick off completed follow-ups via “Follow-up done”.

The reminder is the real value: in most cases, an application without a follow-up goes unanswered, and without a reminder, nobody notices exactly that.

Once a day, Aurora Hive Sales automatically fills the review list with new matches. All you have to do is review and approve.

The automatic run searches and rates — it does not send. Your approval always stands between the automatic run and the outbox.

Bear in mind that every run incurs costs. If the review list fills up faster than you can work through it, that is a reason to slow down the automatic run, not a reason to click faster.

Yes. The ? icon at the top opens the built-in step-by-step explanation of this area — with all the details of the workflow. It can be opened again at any time.

In addition, there is the tour via the button at the top right, which shows the same on the real interface.

Yes — via “View draft” you adjust the subject, text and recipient.

Do this at least for the first dozen. In the process, you learn what the model regularly gets wrong and can improve the text at the root instead of correcting it every time.

They move to the Declined column and stay there. They do not disappear, and that is a good thing: the declines are the proof that every suggestion was looked at by a human.

That is exactly what matters when someone asks how the selection came about — with AI-supported pre-selection, human review is not a courtesy but an obligation.

So that it is traceable in-house what has gone out — applications are commitments made in a person’s name, and nobody should find them only in one individual’s outbox.

During setup, check which address this copy goes to, and that someone actually looks in there.

Yes — and before the application goes out, not afterwards.

Two reasons: legally, an application in a person’s name is processing of their data, which must be covered. In practice, it is the quickest way to lose someone: anyone who finds out about an application because the company calls them will work with someone else afterwards.

So clarify once, as a matter of principle, which kinds of positions you may suggest automatically.

Usually because of the stored profiles, not the search: if the roles and skills of the interim managers are thinly maintained, the rating has nothing to work with.

Work through it from top to bottom: are roles set? Is the availability up to date? Does the region fit? And does the run actually complete — or does it abort because a budget is used up?