LinkedIn outreach with AI

Coco drafts LinkedIn outreach grounded in real context: your prospect's recent activity, mutual connections, role fit, and public commentary. About 4 credits per draft. Connection requests, InMails, and follow-up DMs all queue for your review; nothing sends without your approval at the start. Coco doesn't run spray-and-pray automation that gets accounts flagged by LinkedIn's behavioral detection. That's not the product. It drafts the message, attaches the why-this-prospect context, and hands the send to you (or to approved API access if you've connected it). The result is LinkedIn outreach that reads like you wrote it after homework, because that's what Coco did on your behalf.

Try Coco for free → · 1,000 credits free · no card · ~2-min setup

Why LinkedIn automation gets accounts flagged

LinkedIn has spent years tuning its behavioral detection against bulk-send tools, and it's good at its job. Aggressive sequencing platforms (the ones that promise "send 200 connection requests a day" or "auto-message everyone who views your profile") trip account-level signals that LinkedIn associates with bot behavior. The patterns are well-known: rapid clicks across many profiles in a short window, identical-template messages at volume, message rates that don't match human work-hour cadence, browser fingerprints that match automation libraries.

The downstream cost lands on your account. A warning, then a restriction, then a permanent ban on the profile you've spent years building. The volume those tools promise rarely justifies the account risk for anyone with a real professional reputation tied to the profile.

Coco doesn't operate that way. There's no bulk-send mode, no parallel-account simulation, no race against detection. The model is: Coco drafts, you send. Or, for teams that have connected approved LinkedIn API access, Coco sends inside the strict rate limits LinkedIn allows. Honest about the constraint, instead of pretending the constraint doesn't exist.

How Coco drafts a LinkedIn message

The drafting workflow runs in four steps:

1. Pull the prospect's public signal. Coco reads the prospect's recent LinkedIn activity: posts they wrote, posts they commented on, role changes, content they shared. About 0 credits for the read.
2. Identify the hook. Coco looks for the specific opening that makes the message not generic: a post they wrote last week, a role transition that matches your ICP, a mutual connection that gives you a warm path, a problem they referenced in public.
3. Draft in your voice. Coco writes the message using your voice training (on the Founder tier) or by inferring from your sent history (lighter inference on the Hobby tier). The draft cites the hook directly ("saw your post about CRM data quality last Tuesday"), not generic flattery.
4. Queue for your approval. The draft lands in your Coco queue with the prospect's profile, the hook attribution, the rationale, and the draft itself. You review, edit inline, approve, or send back.

About 4 credits per draft. For a multi-touch DM sequence (connection request → first message after acceptance → follow-up if silent), Coco drafts all three touches up front; the credit cost scales linearly per draft. See how Coco drafts cold outreach by email → for the same drafting pattern applied to Gmail.

Approval, send, and the LinkedIn behavior model

By default, the flow is: Coco drafts → you review → you send manually inside LinkedIn. Coco never operates as a bulk send-tool. The drafts queue up, you work through them on the cadence that matches your normal LinkedIn behavior, and the account stays unflagged because the send pattern is yours, not a script's.

If you've connected approved LinkedIn API access (for Sales Navigator users, this is the standard path), Coco can send drafts on your behalf, but inside the strict rate limits the API enforces. That means roughly the volume a human would do in a day, not the volume an automation tool would claim. The brand promise stays the same: nothing sends without your approval at the start, and the workflow only graduates to autonomous send once you've explicitly authorized that pattern for a specific message type.

This is the same approval-gate model that runs across Coco's follow-up workflow. The discipline is the wedge.

Follow-up sequencing

Most LinkedIn outreach lives or dies on the second touch. The connection request lands, the prospect accepts but doesn't reply, and the conversation goes dormant unless you follow up well. The follow-up is where templated automation tools fall hardest. Every other founder has seen the same "Hey, just bumping this to the top of your inbox!" DM, and the response rate reflects it.

Coco drafts the full sequence in advance with the same context anchoring. The second-touch DM references something specific the prospect did since the connection accept (commented on a post, shared an article). The third touch shifts the angle if signals warrant: maybe to a useful resource, maybe to a different problem hook. Each draft is approval-gated independently. If touch one lands a reply, the rest of the sequence pauses automatically.

About 4 credits per draft across the sequence. A three-touch DM sequence costs roughly 12 credits total, drafted up front.

Try Coco for free → · 1,000 credits free · no card · ~2-min setup

Why this beats LinkedIn automation tools (and beats Lavender for LinkedIn specifically)

The two categories Coco displaces look different but share the same gap.

LinkedIn automation tools (Lemlist, Expandi, Dripify, etc.) promise volume. The trade is account risk and message quality. The tools work until they don't, and when they don't, the cost falls on your profile.

LinkedIn coaching tools like Lavender help you write better, in-line, when you're already at LinkedIn writing a message. They don't draft from scratch; they don't pull context for you; they don't queue a workflow. Lavender is a writing assist; Coco is doing the writing.

Coco sits between the two. It drafts the message from scratch with real context, hands you the send, doesn't trip detection, and scales to the volume your actual LinkedIn behavior supports without faking it. The message quality stays high because the per-draft cost (about 4 credits) is calibrated to make personalization affordable, not so cheap that quality drops to spam.

Frequently asked questions

Will Coco get my LinkedIn flagged?

No, by design. Coco drafts the messages; you send them manually or through approved API access at rate-limit-compliant pacing. There's no spray mode, no parallel-profile automation, no detection-evasion behavior. The send pattern stays yours.

Can Coco send InMails directly?

Yes, if you've approved that workflow and connected the appropriate LinkedIn API access (typically Sales Navigator). Sends still queue for approval by default until you've authorized autonomous send for a specific pattern. Direct sends respect LinkedIn's published rate limits.

What about connection requests?

Drafted with the optional note pre-filled. You approve, you send. For prospects you're sending to in bulk over time, Coco drafts the batch with per-prospect personalization; you work through approvals on your normal cadence rather than at automation speed.

Can I run multi-touch DM sequences?

Yes, all approval-gated. Coco drafts the full sequence in advance (connection request → first DM after accept → follow-up if silent), each touch grounded in fresh context for that prospect. Sequences pause automatically if a touch lands a reply.

Does Coco work for Sales Navigator?

Yes. The LinkedIn integration supports Sales Navigator, and the additional filters and InMail access available there are surfaced inside Coco's drafting and prospecting workflows. The Apollo integration pairs naturally if you're cross-referencing prospect data between the two.

Get started

Connect LinkedIn, hand Coco a prospect or a target list, and see the first drafts. The 1,000 free credits cover roughly 250 LinkedIn drafts on the Hobby tier, enough to evaluate the quality before you commit to a workflow.

Try Coco for free → · 1,000 credits free · no card · ~2-min setup

Or book a walkthrough → if you'd rather see how the drafts come out first.