Outbound

Twin lead-gen engines

A keyword engine plus a LinkedIn engine, from finding buyers to automatic outreach.

What it does

The keyword engine is an autopilot chain: AI extracts keywords from your ICP, searches buyer websites across sources, filters through two gates (qualification plus fit scoring), then generates a personalized multi-step sequence per company — the whole chain runs in the background, pausable and resumable, with live status.

The LinkedIn engine targets decision-makers directly: AI translates your ICP — or just a one-line description of your target customer — into structured filters, paginates the search, fit-scores every prospect, and qualifying ones get an opener generated and outreach started. CRM cross-checks flag already-known and already-won companies, and buying signals weight the queue so the right person is always on top.

Sourcing's next move is outreach, and that chain is wired too: qualified prospects get a personalized email sequence generated in one batch, with human approval as the only outbound gate — enforced server-side, so bypassing the front end still won't send. Prospects who don't reply within a few days, and still score, auto-escalate into a LinkedIn connect, enqueued within a daily budget. LinkedIn runs on a multi-account pool: each account has its own quota, freezes the instant it trips a risk check, and executes fully isolated from the main system. Both channels form their own funnel side by side, so where to lean your effort is a data decision.

Core capabilities

Email + LinkedIn dual-channel

Email first; unanswered mail auto-escalates to a LinkedIn connect. Approval is the only outbound gate, with daily caps and a breaker guarding your domain.

Describe a buyer in one line

Plain-language target descriptions auto-convert to filters; qualified prospects get automatic first outreach.

Permanent global dedupe

Same person, email or LinkedIn profile is never contacted or billed twice across batches.

Feature checklist

  • Dedupe matches any of three identity signals — person ID, normalized LinkedIn URL, normalized email — under a database-level unique constraint
  • Two filter gates: qualification (gate) plus fit scoring; fit ≥ 40 qualifies
  • A one-line target-customer description auto-converts to structured filters; multi-round sourcing keeps only qualified prospects
  • LinkedIn prospects are batch-scored — one call grades a whole batch, an order-of-magnitude faster on fit scoring, with qualified prospects surfacing as they are found
  • Priority = fit ×0.7 + buying signals ×0.3, and smart recommendation surfaces the right person first
  • Audit before you spend: seven range checks catch traps like an all-industry search or contradictory filters, a free preview estimates how many people you'll hit, negative keywords hard-filter, scoring is anchored to S/A/B/C bands, and borderline leads are flagged for human review
  • Contact unlocking runs behind three guardrails — per-user daily quota with atomic reservation, a global credit cap, and a per-contact fit gate for phones; reps self-serve unlock their own prospects, with results backfilled automatically
  • Sequences include locale checks so wording adapts per market
  • Cold outreach passes three eligibility gates — already a customer, already in the inquiry pool, already unsubscribed are all blocked — and only real unlocked/verified emails are sent, never placeholder addresses: existing customers are never disturbed
  • Human approval is the sole outbound gate for cold email, enforced server-side; beyond a daily send cap, a failure-rate breaker auto-halts sending and alerts, so sender-domain reputation is watched
  • Unanswered email auto-escalates to a LinkedIn connect: only fit-qualified prospects consume quota, enqueued within a daily budget, and anyone who replied or unsubscribed is never escalated
  • LinkedIn multi-account pool: per-account weekly-connect and daily-message quotas, connects to one company auto-spread across accounts, and any risk signal freezes the account and suspends its tasks; execution is isolated — the product never holds your LinkedIn session
  • Dual-channel funnel compared per deduped prospect: enqueued → sent → replied → promoted → won, Email and LinkedIn side by side, so resource allocation is data-driven
  • Team-scale prospect management: per-rep daily-new and open-lead caps, auto-assignment favoring the least-loaded rep, and preview-then-reclaim of stale leads; reps see only their own, team leads see their team

Want to see this module run?

Book a demo and we'll run this module on real data.