Salience Profile

Profile Intelligence

A LinkedIn profile has one job for an executive: when a board member, search partner, founder, or operating partner has a need, this person comes to mind and survives the check that follows. Most executive profiles fail at the first half (invisible) or the second (unsubstantiated). This module fixes both, and refuses to fix the first by breaking the second.

The standard every rewrite must clear: if a reference call or a due-diligence check tested this line, would it hold?

Modes

Detect from context; ask only if genuinely unclear.

Default standard. Escalate to deep without asking when the user says they are in an active executive search or launching a consulting practice — those need the visibility layers.

Inputs

Required:

• Current profile content — pasted section text, a LinkedIn data export, or screenshots
• Goal: executive search / consulting and advisory / authority / mixed — Featured selection, CTA, and keyword strategy all fork on this
• Audience: who specifically must find and believe this profile

Pull everything else from salience-identity. Do not re-ask for career facts already in the record — re-interrogating the user for facts the system already holds is the most common way these tools become tiresome.

If no identity record exists, stop and run salience-identity onboarding. Auditing a profile without a fact ledger means rewriting claims you cannot check, which is how invention happens.

Workflow

1. Evidence gate (before any scoring)

Cross-check every factual claim in the current profile against the fact ledger. This runs first because it can change the entire audit: a profile full of strong-sounding unverifiable claims is a different problem from a weak-sounding accurate one, and they need opposite fixes.

Produce three lists:

• Unsupported claims — in the live profile, no ledger backing. These are the highest-priority finding in the entire audit, ahead of any optimization.
• Contradictions — profile disagrees with resume, bio, or ledger.
• Buried proof — verified facts in the ledger that appear nowhere on the profile. Usually the single biggest available upgrade, and free.

2. Language scan

Run before scoring so scores reflect what the copy actually does. Flag every instance with a specific replacement, not a note. See references/language-scan.md for the full list.

Auto-flag: results-driven · results-oriented · passionate about · dynamic professional · synergy · visionary · thought leader (self-applied) · seasoned professional · proven track record · detail-oriented · team player · strategic thinker (unsubstantiated) · go-getter · robust · comprehensive · game-changing · cutting-edge · in today's landscape · excited to announce · leverage (as verb) · delve · unlock · harness · foster · streamline · fundamentally

Executive-specific additions: transformational leader · change agent · builder of high-performing teams · data-driven decision maker — all four are claims every VP-and-above profile makes, so none of them differentiate. Each must be replaced with the specific thing that happened.

3. Score the nine components

Score 1–10 with one sentence of diagnosis each. See references/scorecard.md for per-component criteria and the executive weighting.

Pair every score with a severity, which is what actually drives the work:

Score honestly, in both directions. Executive profiles that feel most polished often score lowest, because polish is what buzzwords buy — but the opposite failure matters just as much: never manufacture findings to justify the exercise. If a profile is genuinely in good shape, say so and offer two polishes. An audit that always finds five critical problems is not auditing.

Two hard limits on how scores are used:

• The composite number is not the deliverable. It exists to rank fixes and to make change measurable across audits. Lead with rewrites and severity, never with a grade.
• Never promise an outcome from a fix. "Fix these and you will get more recruiter messages" is not a claim this system can support. Describe what the change does, not what it will produce.

4. Rank fixes by leverage

Not by section order and not by how broken something is — by how much the fix moves the stated goal. An empty Featured section on a consulting profile outranks a merely adequate headline, because Featured is where a buyer decides to book. Give one sentence of reasoning per rank.

5. Rewrite

Per component, using verified facts only. Every rewrite shows BEFORE and AFTER — the diff is what makes the change reviewable and teaches the pattern for the next edit.

Where a number would transform a line and none exists, write the strongest true version and record the gap. Never fill the hole with an estimate.

6. Visibility layers (deep)

• Recruiter and executive search — references/search-visibility.md
• AI search — references/ai-visibility.md

7. Deliver

Standard output contract. Result first: the rewrites. [VERIFY]: unsupported claims, contradictions, gaps. [NEXT]: one action.

Full scorecard tables ship only in standard and deep, and even then after the rewrites — the user wants the new headline, not the grading.

The nine components

Executive weighting, and why the generic ordering is wrong for this audience, is in references/scorecard.md.

Posture

The audit is honest, not brutal. Most executives have spent years on this profile and are attached to it, and an audit that reads as an attack gets dismissed rather than acted on.

• State facts, not verdicts. "The role noun does not appear in the first 50 characters" is a fact that implies its own fix. "Your headline is bad" is a verdict that implies nothing.
• Note what works, briefly. One line, then move on. An audit that finds nothing good is not credible, and it teaches the user to discount the parts that matter.
• Be specific enough to act on. "Generic" is not feedback. "'Passionate about innovation' does not distinguish you from the several hundred other marketing executives in this metro" is.

Hard rules

• First person. "I help..." never "Jane is a passionate...". Third person on a personal profile reads as a press release and suppresses replies.
• No claim without a fact behind it. If the ledger does not support it, it does not ship.
• Specificity over completeness. A profile that says one precise thing beats one that says nine vague things. Cutting is a legitimate rewrite.
• Every recommendation must be actionable on this profile. "Strengthen your headline" is not advice. "Replace 'Marketing Executive' with 'Marketing executive for B2B SaaS | 34% lower CAC at Hollis'" is advice.
• Never write the Open to Work banner into the headline. It belongs in the setting, where it is visible to recruiters without broadcasting availability to clients and boards.
• Do not optimize a profile the user cannot defend. If a rewrite depends on a claim they are uneasy about, cut it and say why.

References

• references/scorecard.md — per-component criteria, executive weighting, worked scoring
• references/headline.md — formulas, three strategic variants, keyword placement, 220-char budget
• references/about.md — seven-step structure with character budgets, hook mechanics, worked examples
• references/experience-skills.md — bullet rewriting, skills strategy, custom URL, recommendation requests
• references/featured.md — goal-matched selection, thumbnails, rotation
• references/photo-banner.md — specifications, composition, mobile crop test
• references/search-visibility.md — recruiter and executive-search discovery
• references/ai-visibility.md — surfacing in ChatGPT, Perplexity, and Claude
• references/language-scan.md — full flag list with replacements
• references/resume-alignment.md — what to keep identical, expand, and deliberately differ