How to Automate High-Ticket Client Discovery & Qualification as a Solopreneur

The Job to Be Done (Not a Tool Shopping List)

High-ticket discovery is not “build a bigger list.” It is a weekly operating rhythm: identify executives in-market for your offer, verify you can reach them, score fit with evidence, draft outreach you would actually send, and hand off replies to a CRM you will check — before you spend an hour on a discovery call with someone who was never qualified.

Executive search, private banking, enterprise SaaS, boutique consulting — the friction is the same: small addressable market, high cost of a wrong meeting, zero tolerance for generic spray. Tools come after the pipeline is written.

• Success metric: qualified Tier A leads approved and sent per week with ≥ 8% positive reply on first touch.
• Failure metric: bounced emails, hallucinated openers, or calendar filled with unqualified calls.
• Out of scope for v1: multi-channel ads, full SDR team handoffs, automated LinkedIn DMs at scale — scrape narrow, approve sends yourself.

The Discovery Pipeline Map — Scraping → Enrichment → Filter → Route

Eight ordered steps from ICP definition to CRM handoff. Do not skip to enrichment before dedupe — duplicate sends to the same CHRO end deals, not just annoyance.

Network Scraping Layer (Apify & Phantombuster Patterns)

Scraping is the intake valve — not the strategy. Cap batch size, normalize to one schema, and stage before enrichment burns credits on garbage rows.

1. Step 1

   Apify — LinkedIn Sales Navigator Search Scraper

   Schedule actor run with your saved search URL; output dataset to JSON with `fullName`, `title`, `companyName`, `linkedinUrl`, `companyWebsite`.

   Mapping:

   Webhook on run SUCCEEDED → Make/Zapier receives `defaultDatasetId`; HTTP GET dataset items.

   Failure mode:

   Session cookie expired — refresh LinkedIn session in Apify input; never run without `maxItems` cap.

2. Step 2

   Phantombuster — LinkedIn Search Export

   Agent pulls search results into CSV; map to same canonical keys as Apify path for downstream merge.

   Mapping:

   `profileUrl` → `linkedin_url`, `jobTitle` → `title`, `company` → `company_name`.

   Failure mode:

   LinkedIn rate limit — Phantombuster pauses agent; add 15-min delay between batches, not parallel agents.

3. Step 3

   Make — Iterator + Set variable (Normalize row)

   Build `prospect` object per row; slugify company domain from website URL or clearbit-style guess.

   Mapping:

   `prospect.linkedin_url` is dedupe key; strip query params from LinkedIn URLs.

   Failure mode:

   Missing company website — queue row to manual review sheet, do not pass to enrichment (wastes credits).

4. Step 4

   Google Sheets or Airtable — Staging table

   Append-only staging with `batch_id`, `scraped_at`, `source` (apify|phantom). Human can spot-check 5 random rows.

   Mapping:

   Status column: `raw` → `normalized` → `enriched` → `scored` → `approved`.

   Failure mode:

   Duplicate append on webhook retry — dedupe on `linkedin_url` + `batch_id` before insert.

Data Enrichment (Apollo.io & Hunter.io Patterns)

Enrichment turns a LinkedIn row into something you can email. Match rate varies by industry — log it per batch so you know when to fix ICP targeting vs swap providers.

Stale titles common; always pass `linkedin_url` for cross-check in LLM step Weak on subsidiaries and holding companies — parent domain mismatch drops match rate Operational complexity — document fallback order or you double-pay randomly

Semantic Filtering — LLM Qualification Rubric & ICP Scoring

Rules catch obvious junk; the LLM scores nuance (title vs actual remit, timing signals in bio/posts). Weighted rubric keeps scores comparable week over week.

Qualification output (target JSON shape)

{
  "linkedin_url": "https://www.linkedin.com/in/jane-doe-exec",
  "icp_score": 82,
  "tier": "A",
  "why_now": "New CHRO at 400-person fintech; posted about leadership bench 6 days ago.",
  "disqualify_reasons": [],
  "suggested_opener": "Saw your post on building the exec bench — we help CHROs at Series C–D fintechs cut time-to-offer for VP+ roles by 40%.",
  "email_verified": true
}

Use JSON mode in your HTTP module; reject rows where `disqualify_reasons` is non-empty regardless of score. Calibrate on hand-labeled profiles before trusting automated routing.

Personalized Cold Outreach Routing & Human QA Gates

Routing maps score to sequence depth. Every Tier A and B row hits a human approval queue before your ESP sends — automation prepares; you commit.

• Tier A — High conviction

  Score 75–100QA required

  3 touches via Email + LinkedIn connect note (day 3). Personalization: LLM `suggested_opener` + one firmographic detail you verify manually

• Tier B — Worth testing

  Score 50–74QA required

  2 touches via Email only. Personalization: Template with `{company_name}` and `{title}` merge — no fake "loved your post"

• Tier C — Nurture / archive

  Score 0–49

  0 touches via None — tag in CRM for quarterly re-score. Personalization: Do not send cold email; optional newsletter add if opt-in exists elsewhere

Human QA gates

• Email verification pass

  Hunter/Apollo verifier green; no role-based `info@` unless that is normal for segment.

  When: Before Tier A or B enters send queue

• Opener sanity

  Read LLM opener aloud — no hallucinated funding round, podcast, or mutual connection.

  When: Every Tier A send; spot-check 20% of Tier B

• Volume cap

  Max 15 new Tier A contacts/week as solopreneur — protects domain reputation and your calendar.

  When: Weekly review before enabling automation send

• Compliance & opt-out

  Physical address in footer; one-click unsubscribe; no scraped personal emails for consumer B2C.

  When: Before first live batch