Skip to main content
When your own candidates are not enough for a job, Flo can search Google for public LinkedIn profiles that fit it. You see about ten people at a time, and only the people you keep become candidates.

Steps

  1. Open the job and ask Flo, for example “Find hospital COOs in Vietnam outside our database” or “Source NPI project managers at Foxconn or Intel”.
  2. Flo reads the job’s ideal candidate and puts the search on a card in the chat. Each filter (job titles, places, employers, language, words to leave out) is a row of chips, one per value, and each chip says where it came from: “your ask”, or “Flo expanded” when Flo turned a broad word into exact titles. Under “Not used”, everything a search can’t filter on is listed with what happens instead: “Checked on each result” or “Not possible” from search results (for example years of experience, nationality, or what someone did before their current job). If a word could mean very different searches, Flo asks one question first; Skip searches the broader meaning. One line says how many searches Flo will run and the most they cost. If you need something, not just prefer it (for example “from luxury hotels”), say so: Flo marks it required, and people whose search result does not show it are left out. For everyone at one company (“every program manager at VinDynamics”), Flo searches each job title at the company on its own and reads further pages of Google results, so you see about 25 people. The card shows the most this costs.
  3. To change a chip, click it, type the new value and select Save; × takes a value off and + adds one. The card becomes the next version, for example “v2”, with its new cost. You can also tell Flo what to change, and its new card takes the old one’s place. Nothing is searched until you click Run. Saying “go” in the chat does not run it. Cancel closes the card without searching. While the search runs, the card says how many people it has found so far. When it is done, the card folds to one line (what was searched, how many people and what it cost) and the results appear under it. Edit on that line opens the chips again, and Flo plans the next version from your changes.
  4. The search runs on Ascern’s side once you click Run, so a reload, a closed tab or a second tab doesn’t stop it; the card shows where it is. After it runs, the results are a short card in the chat, “Found outside · Financial Controller”, and Open shows them beside the chat as your Candidates table: the same columns as your first Candidates view, filled with each person’s name, headline as their job title, company and place, and their LinkedIn page. Above it is what was checked, for example “3 of 4 verified · title matches, in Vietnam, current role · 1 not confirmed · 6 left out”. Verified people come first: the search result shows the title you asked for, a place you asked for and their current role. A “not confirmed” person has a question mark at the end of the row; hover it to see what the search result doesn’t show, for example “the profile shows no place”. Job pages, people an earlier version already showed, people you marked, roles that have ended, other places and the words you left out are counted under “Left out”, with the reasons. To see about ten more people from the same search, click “More people”. Flo searches further, one job title at a time where Google has no next page, charges only for those searches, and never shows someone the results already show. When no search has more, the results say so and the button goes.
  5. Edit a name or job title if the search read it wrong, tick the people you want and select “Add 3 to Financial Controller · To review” at the top. Verified people start ticked. Only the ticked people are added, with your edits, and each one then reads “Added”. You can also tell Flo who to add.
  6. The thumbs-down at the end of a row marks a person as not a fit for this job. Say why in a few words if you like, such as “a recruiter” or “wrong city”. It never stops you adding anyone. The next search for the same job leaves that person out and says so, and its card lists them under “People you marked”, where × brings them back. A reason that names a kind of profile, such as “a recruiter” or “too junior”, adds that word to Leaves out on the next card, marked “from your feedback”.
When you add someone, they join the job’s To review with the source X-ray. If Ascern already holds that person (the same LinkedIn link), Flo uses the existing candidate instead of adding a second one. If they are already on the job, nothing changes. People marked Do not contact are not added.

What it costs and what it reads

Each search costs a fraction of a cent from your agency’s monthly lookup budget, and the card shows the most a round can cost. Searches are recorded under the person who ran them in Settings › Usage. If the budget runs out, the search provider is slow, or the round stops before it finishes, the card keeps the people it found and says so; click Try again under the table to carry on, and nothing already searched is charged twice. A card waits an hour; after that it says it expired, and Plan again asks Flo for a new one. Search cards stay in the chat and are not listed in Tasks › To review. Flo reads only what Google shows in its search results. It never opens a LinkedIn profile, never signs in to LinkedIn, and never looks up anyone’s email or phone. People you do not keep are deleted from the search after 30 days.