What people hand off on LinkedIn
The LinkedIn jobs dassi users bring us are small and repetitive. One person asked it to go down the Catch up page and send congratulations one at a time, since a two-line note to a former client who just started somewhere new is worth sending but dull to write thirty times a month. Others want the conversation they just had in LinkedIn messages to land in HubSpot or Salesforce without retyping it. Founders who do their own outreach want a sheet of the people from a search who fit, with profile links, to work through later in the week, and recruiters want the same sheet with different columns, which is most of what candidate sourcing and screening covers.
In all of these the reading and typing eat most of an hour, and the decisions take a few minutes at the end.
A worked example: the Monday catch-up
You teach this once, in plain English or by doing it while dassi watches, then save it as a workflow and schedule it for Monday at 8:30. Scheduled runs happen in your own browser, so Chrome has to be open at 8:30 (Edge and Brave work too).
- dassi opens your Catch up page under My Network, signed in as you, and reads this week’s job changes, work anniversaries and birthdays.
- It checks each name against your CRM, or a short “people I actually know” sheet, and drops everyone else.
- For the ones left it drafts a note from the three example messages you gave it at setup. A new job gets a different note than a birthday.
- You see each draft and approve it, edit it, or skip it.
- Approved notes go out one at a time from your account. Anyone who replies gets a dated note on their CRM record.
Most weeks this comes to somewhere between five and fifteen messages.
LinkedIn’s limits, and where dassi stops
LinkedIn caps connection invitations per week. It doesn’t publish the number and it varies by account, although around 100 a week is what most people report running into. Its User Agreement also prohibits scrapers and bots, meaning software that copies profiles, or that adds contacts and sends messages by automated means, and it doesn’t carve out an exception for careful ones.
dassi doesn’t send mass connection requests or bulk-scrape profiles, and no setting turns either on. The sourcing workflow reads search results you already have open and copies the handful of people who match your brief into Google Sheets. Anyone who needs five thousand rows by tomorrow needs a scraper, and we don’t make one.
We don’t know exactly where LinkedIn draws the line for a supervised tool sending ten messages a week from a real person’s account. LinkedIn hasn’t said, so read the terms before you start and decide what you’re comfortable with. dassi’s part in that is modest. It works in your own session at about the speed you would, and nothing goes out that you haven’t read.
What it asks you first
When you teach a job you mark which steps need your OK, and most people flag every send and every CRM write. dassi also stops by itself when something needs a person, like two contacts with the same name, or a “new position” post that reads more like a layoff announcement. If LinkedIn signs you out, dassi uses whatever Chrome’s password manager autofills and waits while you type the 2FA code. The page content it reads goes to the model you picked (Claude, GPT, Gemini or a local one), through dassi’s relay if you’re on included credits.
Where we’d start
With the CRM logging, because it’s the lowest-risk job on this page: nothing gets sent to anyone, and a wrong note takes ten seconds to fix. LinkedIn message to CRM walks through that setup. Once it feels routine, LinkedIn and Reddit pipeline operations adds a daily pass over your saved searches, with outreach drafts held until you’ve read them, and by then you’ll have a better sense than we do of how much of your LinkedIn week you actually want to hand over.