DisperseRuns on your own LinkedIn account
LinkedIn outreach, fully automated. You just approve the send.
Disperse finds your leads, writes the opener, and drafts every follow-up on its own, running on your real LinkedIn session at a human pace. It only writes what it can verify about someone, and nothing goes out until you say so.
No shared datacenter IPs. No mail-merge. One human, one browser, one real account.
Alex RiveraHead of Ops · Series A logisticsVerified match
Qualified
Connect sent
Pending
Connected
Replied
matchShared employer, three years back, verified
draftOpener queued for your review
sentConnection note sent, 9:41am
acceptAccepted in 4 hours
reply“Would love to compare notes on that”
Pace: warm, unhurriedStage 1 of 5
Every send starts as a draft you can edit or reject before it goes anywhere.
Shared-history claims are checked against real profile data, never invented to sound personal.
Runs through your own browser and IP, at human pace. Not a datacenter farm of look-alike bots.
Why replies dry up
Most “personalized” outreach is the same five sentences with a name swapped in.
People can tell. Disperse only states what it can verify, and treats one honest detail as worth more than ten generic ones.
Generic toolTemplate detected in 1.2 seconds. No real signal.
DisperseBuilt from a computed overlap between both real profiles, or left out entirely.
How it works
Four steps, in the order they actually have to happen.
- 01
Pair your LinkedIn
Install the extension and sign in once. From then on, Disperse acts through your real session: your cookies, your IP, your pace.
- 02
Set your voice, once
One core claim, stated exactly how you'd say it. The AI is only allowed to use your words, not improvise new ones.
- 03
Review, or trust the tier
High-value leads route to your inbox for a look before anything sends. Lower tiers can auto-send once you trust the pattern.
- 04
Track every reply
Accepts, replies, and state changes land in one thread, matched to the same pipeline you can watch live.
What's actually enforced
Not settings you could ignore. Guardrails built into how a message gets written.
Human-in-the-loop by default
Every draft is held for review until you say otherwise. If the AI drifts from your stated claim mid-message, the send is flagged and blocked automatically, not just logged.
Verified common ground
Shared schools, employers, and cities are computed from both real profiles. Asserted only when true, silent otherwise.
Seniority-aware tiers
Cadence, tone ceiling, and how directly you ask shift by who you're actually messaging.
Pacing guardrails
Minimum gaps between touches and a hard cap on unanswered follow-ups, enforced before every send.
One inbox, real threads
Replies sync back automatically. No separate tab, no copy-paste between LinkedIn and your CRM.
A sample draft
This is what lands in your review queue.
Names and companies below are illustrative. The checks running underneath are exactly what every real draft goes through before it reaches you.
Core claim, checked verbatim
The highlighted line is the one fact your campaign is built on. If a draft can't include it exactly, it gets flagged, not reworded.
No unfilled placeholders
A draft that still says {first_name} never reaches your queue. It gets caught and regenerated before you ever see it.
Nothing sends on its own
Every opener and follow-up waits here until you edit it, approve it, or reject it.
Follow-up · Alex RiveraAwaiting your review
Hey Alex, thanks for the reply! Since you mentioned the backlog your ops team's been fighting, we built the routing layer three teams like yours are already running on. Worth a 15-minute look, or should I check back in a few weeks?
Core claim intact
Start a campaign
Start a campaignPair your LinkedIn. Set your claim once. Watch the first draft come in.
Free to try on your first hundred leads. No card, no shared browser, no bot farm.