Role intelligence, by consequence.

Every competitor in this category already claims fast shortlists, vetted global talent, timezone overlap and 40 to 80 percent savings. Athyna, Nearshore Business Solutions, Somewhere and Scale Army all say a version of it, and several publish more customer proof than Opus does today. Savings is table stakes, and leading with it invites a price comparison Opus should not want to win.

The defensible territory is the gap between role approved and independent output. Applicant volume is abundant. What is scarce is a small number of people who can own the outcome, exercise judgment, communicate in the buyer's operating context, and use AI in the real workflow rather than listing it as a skill. That is a verification problem, not a sourcing problem, and it is the one Opus is actually built to solve.

The metric we build copy around

Vacancy-to-Productivity Exposure is the value delayed, leadership capacity consumed, and risk accumulated each week from role approval until someone owns the result independently. It beats time-to-fill because a started hire can still be a net drain on the manager. It beats salary savings because it counts blocked output rather than payroll.

Critically, the buyer supplies the business inputs. Public data proves the role exists and what it owns. It cannot reveal what the company is actually losing, so every number we put in front of a prospect stays an editable scenario, never an assertion about their P&L.

Across more than fifty current first-party job postings, the same three consequences recur regardless of function. They are the raw material for every subject line and first sentence we write.

These are compact enough to carry a first line on their own. Each is framed as a hypothesis, never as a claim about what we know is happening inside the account.

Scored on pain intensity, expected conversion and sales efficiency. Read the ranking as a prioritisation tool, not a market size.

Confirmed The modelled revenue column uses a $39,000 first-year value per seat, which matches Opus pricing of a $3,000 placement fee plus roughly $3,000 per month for one dedicated professional. The account pools are reachable-scenario estimates for comparing segments against each other, not counts of real companies.

The PE figure excludes multi-seat and portfolio expansion, so it materially understates what a won sponsor relationship is worth. Track that segment on account value, not per-logo value.

A title starts discovery. What makes an opening worth a candidate-led introduction is the business conversion it owns. We store a normalised family and the raw title, match on synonyms, then score the description separately.

One hundred points, so monitoring does not degrade into title scraping. A campaign fires at 75.

Five public layers, with employer job boards as the anchor because they are the only source that proves work needs an owner right now. Greenhouse, Lever and Ashby all publish their job board data without authentication. Around that we layer SEC Form D filings for recent capital events, FTC merger notices and verified sponsor portfolio pages for acquisition clocks, first-party websites and archived snapshots for operating change, and public GitHub activity for the technical subset.

The gate is two signals. A relevant live role plus one independent corroborator, both with the source URL and observation timestamp preserved. Funding alone is capacity, not pain, and will flood the list with weak accounts if we let it through on its own.

The house style is a short, conversational, candidate-first introduction. That is a deliberate choice: a specific person the reader could interview is more concrete than a diagnosis of their business, and it opens without presuming to know what is broken. The consequence framing does real work, but it belongs on the second and third touch once a conversation exists.

Variables stay visible as literal tokens until the evidence behind them exists. If we cannot fill {CANDIDATE_PROOF} with something true, the email does not send.

For the second and third touches, and for any account where the segment signal is strong, one shared structure with role-specific variables. This is preferable to five unrelated sequences because it gives every campaign one consequence, one transferred-cost detail and one proof-oriented offer.

Every sentence in every campaign carries one of these five states. The operator's last check before send is to name the state of each claim.

Fifteen research artifacts, all generated 2 August 2026 and cited back to primary sources. The interactive Role Intelligence app covers the operational layer: all 25 roles with their monitoring signals and evidence standards, 50 drafted emails, and 10 campaign strategies. Its link and access details are in the covering note rather than on this page.