When you run sourcing through an agent, the data you point it at comes in two kinds. Opted-in data is information candidates entered themselves, on a platform they joined to be found — they own it, they keep it current, and they agreed to be contacted about jobs. Enrichment data is information a vendor inferred about people from outside signals — scraped pages, purchased lists, email pattern matching. The person described never handed it over. That single difference — who owns the data — decides how often your outreach bounces, how often it gets answered, and how much legal and platform risk your workflow carries.
What “opted-in” means in sourcing
A candidate opts in when they create their own profile on a platform where recruiters search, and declare they are open to offers. On Get on Board’s Talent Database, every profile was written by the professional it describes: skills, seniority, country, English level, salary expectation, and whether they are currently listening to offers. Because the profile is how they get their next job, they keep it accurate — the data refreshes itself.
Opted-in data has a second property that matters as much as freshness: consent. When you reach out, the person on the other end asked to be reachable. Your first message starts a conversation, not an intrusion.
How enrichment tools get their data
Enrichment platforms — Apollo, Lusha, ZoomInfo are the usual chain — hold directories of business contacts assembled from third-party signals: scraped public pages, data partnerships, and pattern-matched emails (name.lastname@company.com until proven otherwise). You bring a name you found somewhere else, and the tool guesses how to reach them.
That model was built for B2B sales, and for tech hiring in Latin America it shows three cracks:
- Nobody consented. The engineer in Bogotá whose email Apollo returns never asked to hear about your vacancy.
- The data decays silently. A guessed or scraped address has nobody maintaining it, so it goes stale the day the person changes jobs — and you only find out when the message bounces.
- Regional coverage is thin. These directories skew toward US and European enterprise contacts; LatAm tech profiles are often missing or outdated.
For the full category-by-category comparison, including LinkedIn scrapers, see Tools to reach tech candidates in Latin America.
The difference at a glance
| Opted-in data | Enrichment data | |
|---|---|---|
| Who owns the data | The candidate — they wrote and maintain it | A vendor — inferred about people who never supplied it |
| Consent to be contacted | Yes, declared on the profile | No |
| Freshness | Self-updating; the profile is the candidate’s job search | Decays until a bounce reveals it |
| Reply behavior | Higher — the person expects recruiter contact | Lower — cold contact from guessed channels |
| LatAm tech coverage | The platform’s specialty | Thin, US/Europe-skewed |
| Risk | Consent is on the record; you still handle personal data under your own obligations | Consent gaps, data-protection exposure, scraping fragility |
One layer vs. a chain of services
The difference gets sharper once an agent is doing the work. An enrichment-based agentic workflow is a chain: one tool to discover names, an enrichment API to guess contacts, a verification service to filter bounces, an outreach tool to send. Each link has its own account, its own billing, its own failure mode — and the agent multiplies every weakness at machine speed.
An opted-in source can collapse that chain into one layer. Get on Board exposes an authenticated MCP server at https://www.getonbrd.com/mcp: you paste the URL into Claude, ChatGPT, Codex, or Cursor, sign in, and your agent can search Talent Database from a plain-language brief, read privacy-preserving candidate profiles, and pull the description of your open jobs to source against. No verification step, because there is nothing inferred to verify.
The privacy model is part of the design: your agent never receives emails, phone numbers, CVs, or social links — contact data stays gated. When the shortlist is ready, you open each profile in the web app and unlock contact details there, with credits, like always. The agent narrows; you decide who to contact.
Where enrichment tools still fit
Enrichment platforms remain the right call when the person you want has not opted in anywhere — sales prospects, partnership targets, executives you are headhunting outside any talent pool. For those, inference is the only option. For tech hiring in Latin America, where an opted-in pool exists for exactly the people you want, paying to guess data that candidates already volunteered elsewhere is the expensive path to a worse reply rate.
More from Talent database and sourcing
Sign in with Google
