Use LinkedIn Search as a Qualification Workflow

Write the hypothesis before opening search

If you begin with 'VP Sales,' LinkedIn will give you people who share a title and little else. Begin with a sentence you can test: companies of this kind, experiencing this change, may need this outcome, and this role is likely to care because of this responsibility. Can you say what would disqualify the company? If not, your search is still a market category, not a buyer hypothesis. OKKI Go can hold a company-search hypothesis before LinkedIn role research begins, keeping account selection separate from profile interpretation. Search accounts first, then roles. Recheck current Sales Navigator filter names on the official help page; a title is a clue, not a buyer.

Define geography, company type, size only where it matters, relevant technologies or operating conditions, and exclusions. Then describe the role in terms of work rather than a single title. You may need the problem owner, technical evaluator, budget holder, or implementation partner. Different organizations name those jobs differently.

Use current filters as research controls

LinkedIn and Sales Navigator controls change, so this platform guidance was checked on August 17, 2026. Verify current official documentation and your account's available interface. Use account filters to narrow the company pool, then lead filters to explore roles inside retained accounts. Why separate them? Because a plausible role at an unsuitable company is still a poor prospect.

Use inclusion and exclusion deliberately. Save the search only after reviewing false positives, then note which criterion produced them. A keyword may describe a past job, a personal interest, or an unrelated product. You should read the profile and company context before carrying that term into your CRM as a fact.

Look for change without inventing intent

Company posts, leadership changes, hiring, product launches, expansion, and professional activity may give you timing context. Ask what the event actually establishes. A hiring post can show investment in a function; it does not prove a budget for your offer. A person's post can reveal a concern; it does not make private life or unrelated activity fair material for personalization.

Keep the source and date. Would the observation still support your message six months later? If timing is the entire reason for contact, stale evidence should return the record to research.

Map a buying group, not a heroic decision-maker

You may find several plausible people: an operational owner, a technical evaluator, procurement, finance, and an executive sponsor. You do not need to message all of them. You do need to understand why each role might enter the decision and who can correct your hypothesis.

Ask yourself: if this person replied 'not me,' would you know where to look next? If the answer is no, research the work around the problem. Connections and mutual contacts can improve context, but they do not create permission to imply endorsement.

Decide whether a message is earned

A message is warranted when you can explain the account fit, the role relevance, and the observation that makes the question timely. Write the first note to test that reasoning. State the business context, make one bounded connection to the recipient's role, and offer an easy correction or next step. You are not announcing that LinkedIn has identified a buyer.

If the evidence is weak, follow the company, refine the search, or wait. Restraint is part of prospecting skill because every premature message consumes attention and makes later contact harder.

Move evidence into a controlled workflow

Do not copy a profile into a list without provenance. Record the URL or source, observation date, company reason, role hypothesis, trigger, reviewer, and permitted next action. When the profile changes, you can distinguish a new fact from an old assumption.

OKKI Go can support a separate company-search and review path, including natural-language company criteria, candidate correction, contact discovery, draft preparation from approved context, and user confirmation before sending. Use it to keep the reasoning inspectable. It does not turn LinkedIn profile data into verified need or remove your obligations for lawful, relevant outreach.

• Account search iteration: Write the exact account conditions and exclusions before choosing filters. Review the first results and label why each false positive appeared. Change only one filter or keyword, rerun the search, and preserve the better version. Can another researcher explain what the adjustment repaired?
• Role search iteration: Describe the work connected to the problem, then test several title families rather than one favored title. Read profiles for current responsibilities and organizational context. Exclude people whose matching keyword belongs to an old position, unrelated project, or personal interest.
• Current-change review: Choose one company event and record its source date. Write what it establishes, what you infer, and the alternate explanation. Your message may test the inference, but it should never present that inference as a confirmed buying initiative.
• Public-detail review: Highlight every personalized detail in a draft. For each one, ask whether it is accurate, current, business-relevant, and proportionate. Remove private-life references, decorative flattery, and facts that cannot explain why this commercial question belongs with this recipient.
• Connection-path review: A mutual contact or shared group can provide context, but you should not imply endorsement. Verify what the relationship actually permits, decide whether an introduction is appropriate, and make it easy for both the intermediary and recipient to decline.
• Buying-group sketch: Place the problem owner, evaluator, procurement role, approver, and implementation owner around the decision. You may not find every person on LinkedIn. Record uncertainty and choose the one contact most able to correct the hypothesis instead of broadcasting to the entire group.
• Message preflight: Can you explain the company reason, the role reason, and the timely observation in three separate notes? Does your opening ask one answerable question? If any reason depends on an unsupported profile inference, return to research before you request attention.
• Platform-date check: Open current LinkedIn or Sales Navigator documentation before training a team on a filter, saved search, alert, or account control. Record the verification date and available account context. The feature view observed on August 17, 2026 should not be treated as permanent.
• Outcome correction: Classify a reply by the hypothesis it changes: account fit, role, timing, problem, channel, or permission. Update the relevant rule and retain the previous version. Your LinkedIn work improves when corrections alter the next search, not when replies merely increase an activity total.
• Quarterly cleanup: Revisit saved searches, retained accounts, role assumptions, and stale triggers. Remove records that no longer meet the written case and document why. A smaller current pool is more useful than a growing archive whose original context has expired.
• Company-page pass: Read the current company description, selected posts, jobs, product pages, and credible external sources for operating context. You are looking for a reason the account matches your hypothesis, not a sentence to flatter. Record the exact source and date, then write an alternative interpretation so your researcher can see how much uncertainty remains.
• Profile pass: Confirm the person's current company, role, tenure only where relevant, stated responsibilities, and professional topics. Your note should distinguish a responsibility the profile states from one you infer. If your entire contact rationale depends on the inference, ask what other source or lower-commitment step could test it before your message becomes assertive.
• Keyword pass: A matching term may appear in a headline, role description, post, past employment, certification, or recommendation. Those locations mean different things. Inspect where your keyword occurs and whether it describes current work. Exclude your decorative matches instead of rewarding them with a relevance score that no reviewer can explain.
• Account-list pass: Review saved accounts for duplicates, subsidiaries, inactive entities, customers, competitors, partners, and excluded geographies. Your result set should reflect the commercial unit that can plausibly buy, not every profile attached to a famous parent brand. Preserve your merge or exclusion decision because ownership structures and territories can alter who actually controls the decision.
• Role-transition pass: A newly appointed leader may be relevant, but your transition story is still a hypothesis. Check what changed in scope, what the organization is doing, and whether your offer relates to an immediate responsibility. Avoid congratulatory personalization that uses the promotion as decoration while your commercial reason remains generic.
• Messaging permission pass: Identify the applicable business context, jurisdiction, recipient type, data source, channel, objection status, and suppression record. Your public-profile research does not settle every requirement. Use current policy and qualified advice where needed, and keep your outreach decision separate from the technical ability to send a message.
• Search-to-CRM pass: Before exporting or entering a record, require your source, observation date, account reason, role reason, timing reason, and status. If a field is inferred, label it. Your future teammate should not mistake a research guess for a verified profile fact simply because both appear in the same CRM column.
• Learning pass: Compare your saved-search criteria with referrals, corrections, and disqualifications. Which filter created your most persistent false positive? Which account condition produced useful conversations across several titles? Change your search deliberately and preserve the previous version, so your next LinkedIn session begins with accumulated evidence rather than a larger but equally uncertain result set.
• Research queue design: Separate accounts ready for role research from accounts waiting for a dated trigger and profiles waiting for verification. Your queue should describe uncertainty, not seller enthusiasm. Set an expiry date for observations and a reason for returning each record. This prevents a saved LinkedIn search from turning into an endless contact backlog whose original commercial context has disappeared.
• Recipient-view test: Hide your internal notes and read only the proposed message. Can your recipient understand why the question belongs with their current work? Can they correct your premise without accepting a meeting? Does your wording expose which observation came from LinkedIn and which connection was your inference? Revise until the burden of interpretation stays with you rather than the person you hope will reply.
• Team-boundary test: Document which LinkedIn actions your organization permits, which data may enter which system, how objections and suppression are honored, and when a record must be deleted or reviewed. Your process should follow current platform terms and applicable requirements. Access to a search result is a technical fact; permission to collect, store, enrich, or message is a separate organizational decision.
• False-positive clinic: Take ten profiles your search retained and hide their names. Ask a teammate to judge the account case, role case, timing case, and permitted next action from your evidence. Compare disagreements, trace each one to a filter or inference, and rewrite your search note. This review tells you whether your LinkedIn method is repeatable or merely familiar to its creator.